Three Studies in Epicurean Cosmology

May 31, 2018 | Author: Anonymous | Category: Documents
Share Embed

Short Description

Jun 26, 2007 - Proefschrift ter verkrijging van de graad van doctor aan de Universiteit Utrecht ...... In his De caelo I...



Copyright © 2010 by F.A. Bakker This work is licensed under the Creative Commons Attribution 3.0 Unported License. For the text of this license, see Cover and book design: F.A. Bakker Cover pictures: Above: bust of Epicurus, Roman copy after Hellenistic original from ca. 200 BC, British Museum, detail from picture by Marie-Lan Nguyen 2006 / Wikimedia Commons ( Below: detail of woodcut from Camille Flammarion’s L’atmosphère: météorologie populaire, Paris 1888, p.163; coloured by Hugo Heikenwaelder 1998; modified by Jürgen Kummer 2010 / Wikimedia Commons ( Printed by Wöhrmann Printing Service, Zutphen. ISBN 978 90 393 5456 8

Three Studies in Epicurean Cosmology Drie Studies in Epicureïsche Kosmologie (met een samenvatting in het Nederlands)

Proefschrift ter verkrijging van de graad van doctor aan de Universiteit Utrecht op gezag van de rector magnificus, prof.dr. J.C. Stoof, ingevolge het besluit van het college voor promoties in het openbaar te verdedigen op woensdag 17 november 2010 des middags te 12.45 uur door Fredericus Antonius Bakker geboren op 19 mei 1971 te Amsterdam

Promotor: Prof.dr. K.A. Algra

For Ruthie


x xii






Introduction......................................................................................................8 Variations in the use of multiple explanations ..................................................9 Truth, probability and personal preferences ....................................................11 1.3.1 Introduction ...........................................................................................11 1.3.2 Are all alternative explanations true?.....................................................12 1.3.3 Contestation...........................................................................................30 1.3.4 Non-contestation and analogy................................................................32 1.3.5 Degrees of probability and personal preferences....................................35 1.3.6 Lucretius’ supposed preference for the theories of the mathematical astronomers...........................................................................................39 1.4 Multiple explanations and doxography ...........................................................53 1.5 The sources of the method of multiple explanations .......................................57 1.5.1 Introduction ...........................................................................................57 1.5.2 Democritus.............................................................................................58 1.5.3 Aristotle .................................................................................................60 1.5.4 Theophrastus..........................................................................................61 1.5.5 The Syriac meteorology..........................................................................64 1.5.6 Conclusions concerning the origins of multiple explanations..................67 1.6 Conclusions....................................................................................................68 2



Introduction....................................................................................................70 Range, delimitation and subdivisions of meteorology.....................................71 2.2.1 Introduction ...........................................................................................71 2.2.2 Ancient meteorological texts ..................................................................72 2.2.3 The table ................................................................................................89 2.2.4 Some observations..................................................................................92 2.2.5 Some conclusions ...................................................................................95 2.3 Terrestrial phenomena other than earthquakes ................................................99 2.3.1 Lucretius ................................................................................................99 2.3.2 Local marvels in meteorology and paradoxography ............................. 102 2.3.3 Conclusion ........................................................................................... 113 2.4 Order of subjects .......................................................................................... 115 2.4.1 Introduction ......................................................................................... 115


TABLE OF CONTENTS 2.4.2 The table .............................................................................................. 115 2.4.3 Some observations................................................................................ 118 2.4.4 The original order of subjects............................................................... 124 2.4.5 Deviations from the original order ....................................................... 125 2.4.6 The internal structure of chapters and sections .................................... 128 2.5 Relations between the four texts................................................................... 130 2.5.1 Epicurus’ Letter to Pythocles and his “other meteorology”.................. 131 2.5.2 Lucretius DRN VI and Epicurus’ “other meteorology” ........................ 131 2.5.3 Authorship and identity of the Syriac meteorology................................ 132 2.5.4 Lucretius, Epicurus and the Syriac meteorology................................... 140 2.5.5 Aëtius’ Placita and Theophrastus’ Physical Opinions .......................... 141 2.5.6 Summary .............................................................................................. 142 2.6 Conclusions.................................................................................................. 144 3



3.1 3.2

Introduction.................................................................................................. 146 Some preliminaries....................................................................................... 149 3.2.1 The shape of the earth in antiquity: a historical overview..................... 149 3.2.2 Ancient proofs of the earth’s sphericity ................................................ 154 3.2.3 Epicurus’ ancient critics ...................................................................... 158 3.2.4 The direction of natural motion and the shape of the earth ................... 160 3.3 Discussion of relevant passages.................................................................... 162 3.3.1 The rejection of centrifocal natural motion (DRN I 1052ff) .................. 163 3.3.2 Downward motion (DRN II 62-250) ..................................................... 190 3.3.3 The apparent proximity of the sun (DRN IV 404-13)............................. 198 3.3.4 Climatic zones? (DRN V 204-5) ........................................................... 199 3.3.5 Lucretius’ cosmogony (DRN V 449-508) .............................................. 202 3.3.6 Stability of the earth (DRN V 534-63)................................................... 213 3.3.7 The size of the sun (DRN V 564-591).................................................... 213 3.3.8 Centrifocal terminology (DRN V 621-36) ............................................. 217 3.3.9 Sunrise and sunset (DRN V 650-79) ..................................................... 218 3.3.10 The earth’s conical shadow (DRN V 762-70) ....................................... 219 3.3.11 The ‘limp’ of the cosmic axis (DRN VI 1107) ....................................... 222 3.3.12 Philodemus and the gnomon (Phil. De sign. xxx 20-27)........................ 230 3.4 Conclusions.................................................................................................. 231 GENERAL CONCLUSIONS


















Tables: Table 1-1: Place of the mathematical astronomers’ explanations in the astronomical section of DRN .................................................................... 43 Table 1-2: Place of the astronomers’ explanations in the astronomical passages of the Letter to Pythocles .............................................................................. 51 Table 2-1: Range of subjects and subdivisions in various ancient accounts of meteorology ............................................................................................. 91 Table 2-2: Lucretius’ account of terrestrial phenomena............................................. 100 Table 2-3: The subjects of DRN VI 608ff with parallels in meteorology and paradoxography...................................................................................... 106 Table 2-4. The order of subjects in Aëtius, the Syriac meteorology, Lucretius and Epicurus ................................................................................................. 117 Table 2-5. Terrestrial subjects other than earthquakes in Aëtius and Lucretius .......... 124 Table 2-6. Proposed original order of subjects .......................................................... 125 Table 2-7. Syriac meteorology chapter 6 and Lucretius VI 219-422 on thunderbolts............................................................................................ 129 Table 2-8. The alternative explanations of thunder in the Syriac meteorology and Lucretius VI ........................................................................................... 130 Table 3-1: Ancient proofs of the earth’s sphericity.................................................... 158 Table 3-2: Passages to be discussed .......................................................................... 163

Illustrations: Figure 2-1: Possible relations between Epicurus, Lucretius and the Syriac meteorology............................................................................................ 141 Figure 2-2: Possible relations between Aëtius, Epicurus, Lucretius and the Syriac meteorology............................................................................................ 143 Figure 3-1: Time-line of ancient theories on the shape of the earth............................ 154 Figure 3-2: Linear or parallel cosmology.................................................................. 162 Figure 3-3: Centrifocal cosmology............................................................................ 162 Figure 3-4: The three climatic regions....................................................................... 202 Figure 3-5: The five climatic zones........................................................................... 202 Figure 3-6: Furley’s calculation of the sun’s distance on the assumption of a flat earth ....................................................................................................... 216 Figure 3-7: The moon in the conical shadow of the earth .......................................... 220 Figure 3-8: Lucr. VI 1106-13: five places, four winds............................................... 228

ACKNOWLEDGEMENTS Having reached, with this dissertation, the end of a long-running project, I would like to acknowledge my debt and gratitude to a number of people who have helped and supported me along the way. The first I would like to mention is Keimpe Algra. I thank him for his invitation, all those years ago, to apply for the position of PhD student at Utrecht University, for the freedom he allowed me to conduct my research as I saw fit, for his always benevolent criticism of my work, and for his never failing trust in me and this project. I also thank Teun Tieleman and Jaap Mansfeld for the interest they have always shown in my work and Jaap in particular for his useful remarks on the first complete draft of my dissertation. My time as a PhD student at Utrecht University has been a happy one, not least because of the camaraderie (as Claartje fittingly called it) of my fellow PhD-students, Maarten van Houte, Claartje van Sijl, Irene Conradie and Albert Joosse, and my roommate for many years, Marnix Hoekstra, with whom I have had so many pleasant conversations about my research and his. In particular I want to thank Irene and Claartje, who, having finished their dissertations, graciously helped me prepare mine for printing. I have also greatly enjoyed, and benefitted from, the many educational activities and occasions to present my research that were organized by OIKOS, the National Research School in Classical Studies in the Netherlands, and particularly the OIKOS Masterclass in Athens, 2006, which I helped organize together with Jacqueline Klooster, Mariska Leunissen and Carolien Trieschnigg. A word of thanks is also due to my colleagues at Werenfridus Highschool in Hoorn, especially Alwies Cock and Machteld Wit, for their continual encouragement to pursue my research and finish my dissertation. Highschool is not, however, the best environment to be working on a PhD. I am very grateful, therefore, to have been given the opportunity to work as an assistant professor at the Free University (VU) in Amsterdam, a position which has proved very congenial to my research, not least because of the support and interest shown by my current colleagues, notably Wouter Goris and Marije Martijn. I am also grateful to the participants in the Research Group Ancient and Medieval Philosophy for their useful comments on a draft version of my first chapter. Yet, not much would have come of my research without the support and warm embrace of my friends and family. From the time when we were both PhD students in Utrecht Joost has been an indispensable friend, even at the present regrettable distance, and I am very glad to have him for a paranimf at my promotion. Robbert, my long-time rowing mate, who has turned out to be



just as indispensable and practically a family member, I thank for being a most loyal and ever helpful friend. Miriam I thank for being the best sister I could wish for, and, as for my brother, I am proud to have Alexander standing by me as my other paranimf. I thank my parents for their enthousiastic support for my research, which means a lot to me. Finally and most importantly, I want to thank Ruth, my wife, for sharing my life and caring for me while I was caring for my dissertation, and our children, Juniper and Thorin, for constantly reminding me that in the end life is more precious than a dissertation could ever be. I want to conclude by briefly commemorating two people I would rather have thanked in person. One is my friend Guus van der Kraan, with whom I studied Greek and Latin at the University of Amsterdam, and who subsequently, having completed his studies with incredible speed and honours, preceded me in obtaining a position as PhD student, but then lost faith in his project and in life itself. I often miss him. The other one is René Veenman, who was my colleague at Werenfridus Highschool for just one year, and whose friendship I was hoping to cultivate as soon as the work on my dissertation was over. At the presentation of his acclaimed book De klassieke traditie in de Lage Landen (‘The classical tradition in the Low Countries’), just a few months before his unexpected and untimely death in March of this year, he gave me a free copy of his work, asking only for a copy of my dissertation in return.

ABBREVIATIONS In this work the following abbreviations will be used: Arr. B-L D-K D.L. E-K FHS&G

M&R 1 M&R 2


G. Arrighetti, Epicuro, Opere, Torino 1960, revised 1973. J. Bollack & A. Laks, Epicure à Pythoclès: Sur la cosmologie et les phénomènes météorologiques, Cahiers de Philologie de Lille, vol.3, Villeneuve d’Ascq 1978. H. Diels / W. Kranz, Fragmente der Vorsokratiker, 3 vols., Berlin 196010. Diogenes Laërtius, Lives of the philosophers. L. Edelstein & I.G. Kidd, Posidonius, I. The Fragments, Cambridge 1972, 19892. W.W. Fortenbaugh, P.M. Huby, R.W. Sharples & D. Gutas (eds.), Theophrastus of Eresus: Sources for his life, writings, thought and influence, Pt. I, Philosophia antiqua Vol. LIV, 1, Leiden 1992. J. Mansfeld & D.T. Runia (1997), Aëtiana, the Method and Intellectual Context of a Doxographer, vol. 1: The Sources, Philosophia Antiqua, vol. LXXIII, Leiden 1997. J. Mansfeld & D.T. Runia (2009a), Aëtiana, the Method and Intellectual Context of a Doxographer, vol. 2 (in two parts): The Compendium, Philosophia Antiqua, vol. CXIV, Leiden 2009. Oxford Latin Dictionary, Oxford 1968. J. von Arnim, Stoicorum Veterum Fragmenta, 3 vols., Leipzig 1903-5; vol.4 with indices by M. Adler, Leipzig 1924. Thesaurus Linguae Latinae, Munich 1894-present. H. Usener, Epicurea, Leipzig 1887, repr. Rome 1963.

“A cosmos is a circumscribed portion of sky, containing heavenly bodies and an earth and all the phenomena, ….” - Epicurus, Letter to Pythocles 3 [88].1

“The cosmos is a system composed of heaven, earth and the natures contained in these.” - Chrysippus, SVF II 527 (Ar. Did. 31).2

GENERAL INTRODUCTION By ‘cosmos’ (kÒsmow: literally ‘order’, ‘ornament’) the ancient Greeks denoted the orderly whole comprising everything within the confines of the outer heavens, or that part of the heavens to which the fixed stars were supposed to be attached. To some, like the Stoics, our cosmos was unique and included everything there is; others, like the Epicureans, believed that there were innumerable such cosmoi beyond the limits of the outer heavens. The word ‘cosmology’, coined by Christian Wolff in 1731 to refer to the study of the cosmos,3 is often used retrospectively to denote that branch of ancient philosophy which occupied itself with the cosmos. In view of the definition given above, it might be thought that ancient cosmology dealt with everything there is, but in practice ancient writings purporting to deal with the cosmos were more limited in scope. Ancient cosmology dealt with the cosmos as a whole, its origin and arrangement and the positions of its major parts. Very often the study of the cosmos blended into a more detailed study of its outer portion, discussing the nature and appearance and the various motions of the planets and fixed stars, and all those phenomena we would call astronomical. Sometimes cosmology also included phenomena closer to home, such as rain and thunder and wind, which we would call meteorological, and finally it might even include phenomena on and below the surface of the earth, such as earthquakes, volcanic eruptions, springs, rivers, the sea. This is true of the only ancient work which explicitly claims to be dealing with the cosmos: Pseudo-Aristotle’s On the cosmos discusses astronomy, meteorology and 1

KÒsmow §st‹ perioxÆ tiw oÈranoË, êstra te ka‹ g∞n ka‹ pãnta tå fainÒmena peri°xousa, ... 2 KÒsmow §st‹ sÊsthma §j oÈranoË ka‹ g∞w ka‹ t«n §n toÊtoiw fÊsevn. The same definition in Posidonius fr. 14 E-K; Pseudo-Aristotle De mundo 391b9-10; Cleomedes I 1, 3-10; etc. 3 Christian Wolff, Cosmologia generalis, Frankfurt / Leipzig 1731.



geography.4 Much the same applies to Epicurus’ Letter to Pythocles: although its subject matter is explicitly stated as ‘lofty things’ (met°vra), the actual discussion starts with the definition of ‘cosmos’ quoted above, followed by a few other cosmic questions, and then moves on – without indicating any break – to astronomical and meteorological matters. What all these phenomena – astronomical, meteorological and geographical – have in common is that they are both observable and non-biological. Ancient cosmology then, may be defined as the study of the cosmos as a whole as well as all phenomena contained in it, insofar as these are observable and non-biological. This dissertation consists of three more or less independent studies of various aspects of Epicurean cosmology, which appeared to me to be in need of further clarification or re-examination in the light of recent insights or – conversely – on account of the uncritical persistence of old interpretations. The emphasis in all three studies will be on historical interpretation: to understand Epicurean cosmological theories in their own historical context. The first study concerns the Epicurean use of multiple alternative theories to account for astronomical and meteorological phenomena. The second deals with the range and order of subjects in Epicurean meteorology as compared to other ancient meteorological texts. The third and final study is concerned with the Epicurean view of the shape of the earth. Before introducing each of the three studies in some more detail, I will first sketch the outlines of Epicureanism, dealing briefly, first with Epicurus himself, then with a number of later Epicureans, such as Lucretius, and their value as sources for a reconstruction of Epicurean theory, and finally with the most relevant parts of Epicurus’ philosophy. Circa 305 BC, Epicurus (341-270 BC) established a new school of philosophy around his newly acquired estate in Athens. Like the competing school of the Stoics, founded around the same time, Epicurus and his followers were principally concerned with ethics, to which other branches of philosophy, such as physics, were subordinate. According to Epicurus, the life of man was oppressed by unfounded fears – of the gods, of illness, of death – which stood in the way of real happiness. It was in order to eradicate these fears and replace them with a true understanding of nature that physics came in, which was largely based on the atomism of Democritus (ca. 460-370). Yet,


On the probable inauthenticity of On the cosmos (De mundo) see n.204 on p.76 below.



despite its ultimately subservient role, physics was no small matter to Epicurus, who devoted the 37 books (!) of his On nature to it.5 Today the bulk of Epicurus’ writings is lost. Only three works have come down more or less complete, having been quoted as part of the account of Epicureanism in book X of Diogenes Laërtius’ Lives of the philosophers. These three are the Letter to Herodotus, a summary of Epicurean physics; the Letter to Pythocles, a summary of Epicurean cosmology (in the sense indicated above); and the Letter to Menoeceus, an overview of Epicurean ethics. Of his other works only scorched fragments – dug up in Herculaneum, where they were buried during the famous eruption of Vesuvius in 79 AD – and a large number of quotations and paraphrases by later authors remain. The fullest account of Epicurean physics we possess today is the Latin didactic poem De rerum natura, ‘On the nature of things’, by the Roman Epicurean Lucretius (ca. 99-55 BC). In six books,5 and over 7,000 verses, Lucretius deals with every aspect of the physical world, from the invisible smallness of atoms to the infinite extent of the universe, and from sex and conception to inevitable death, which he diversifies with ethical exhortations and eulogies of Epicurus. Other Epicureans whose works we can still partly read are Philodemus of Gadara and Diogenes of Oenoanda. Philodemus, a Greek philosopher roughly contemporary with Lucretius, was based in Herculaneum in southern Italy, where many of his works, scorched and buried by the eruption of Vesuvius in 79 AD, were found during excavations in the 1750s, eventually resulting in a number of fairly readable, though fragmentary, editions.6 Diogenes of Oenoanda was a wealthy Greek from Oenoanda in Lycia (southern Turkey), living in the second century AD, who in his old age erected a wall in the town centre which he inscribed with a summary of Epicurean philosophy, many fragments of which can still be read and are available in modern editons.7 Between Epicurus and these three followers considerable time had elapsed and it is only natural to ask whether their teachings were still the same as Epicurus’. Especially Lucretius has become the object of an ongoing debate, in which one side detects traces of later intellectual and philosophical developments throughout his poem, while the other camp with equal vigour interprets any such sign as a reference to philosophical positions that would have been relevant to Epicurus himself. In this work I will try to avoid both


‘Book’ is the standard translation of Greek ≤ b¤blow / tÚ bibl¤on and Latin liber, i.e. a scroll, the size of which may be compared to a chapter of a modern book. 6 Works to which I will be referring are Philodemus Per‹ ye«n (On the gods, ed. Diels, 1916-7) and Per‹ shme¤vn (On signs, ed. E. & P. De Lacy, 1978). 7 See now Smith (1993) and (2003).



extremes. Lucretius’ explicit claim to be following Epicurus’ writings8 should deter us from looking for later developments when we can do without them, but on the other hand we should not reject an obvious interpretation because it might conflict with Lucretius’ supposed intellectual isolation. The basic tenets of Epicurean physics are not difficult to summarize, and here at least Lucretius is in complete agreement with Epicurus. According to the Epicureans there are two basic entities – bodies and space, while everything else is either a property of these or non-existent. Bodies are either compounds or atoms. The latter are imperceptibly small, indivisible and incompressible particles out of which compounds are made and into which they are eventually dissolved again. Among compounds Epicurus also numbers souls and the gods. There are many types of atoms, differing from each other in shape, size and weight. Beside these, atoms possess none of the qualities that belong to perceptible bodies, such as colour, temperature or smell. The number of atoms and the extent of space are infinite. Atoms are forever moving at a constant but inconceivable speed. Left to their own devices the atoms move either downwards by the force of their own weight or they swerve ever so little from their course, but if they collide with other atoms they may rebound in any direction, while in compounds they are reduced to vibrating. Every compound body is constantly shedding imperceptibly thin membranes which preserve its outward shape and texture. When these membranes, which are called ‘images’ (eidōla), enter our eye we see the compound body. The Epicureans also distinguished a kind of ‘mental perception’ which is transmitted by still finer images that impinge directly on the mind, and there produce a kind of ‘mental picture’ which we call a memory, a dream or a thought. Other forms of sense-perception, such as hearing and smell, are brought on in a similar fashion by other kinds of effluences. Although the images and these other effluences are capable, because of their extreme subtlety, to travel almost unimpeded, they can become confused. Yet, by itself every perception is true, because it accurately reports the way it is affected by the external object. It is only when we start interpreting our perceptions and form opinions about the external object that falsehood may arise. For instance, when an oar standing in the water appears broken to us, this perception is true in the sense that the images really convey this impression to us, but when we infer from this that the oar is really broken, this inference may well be false. The only way to find out for sure is by making another observation (which in itself is no more true than the other one) under circumstances where we know impressions to generally correspond to 8

Lucr. DRN III 9-12, quoted on p.131, n.331 below. See also DRN V 55-6.



their objects. We might, for instance, handle the oar, or pull it out of the water and look again. Starting from these basic tenets Epicurus constructed a complex theory that could account for every aspect of reality, and throughout Epicurus’ Letters to Herodotus and Pythocles and throughout Lucretius’ De rerum natura we see these principles being put to work to explain specific phenomena. An important class of such phenomena are the so-called meteōra or ‘lofty things’, to which Epicurus devoted his Letter to Pythocles, and which comprise both astronomical and meteorological phenomena. Lucretius too deals with these phenomena in books V and VI of his De rerum natura. A characteristic feature of these meteōra is that they can only be observed from afar and do not admit of more reliable observations. In these circumstances it is impossible to falsify our hypotheses about them, and we are forced to accept every theory that does not somehow conflict with the basic tenets of Epicureanism. Far from deploring this conclusion, Epicurus and Lucretius blithely embrace it, accounting for almost every astronomical and meteorological phenomenon with a list of alternative explanations. In the first chapter of this dissertation I will explore some general aspects of this method of multiple explanations. In the first place there is the epistemological point of view. Nobody will object to the view that a theory that cannot be falsified must be held possible. There are indications, however, that Epicurus went further than that and claimed that any such theory is true. This would imply that in these cases several, sometimes conflicting, truths exist side by side. The evidence for asscribing this position to Epicurus is not unambiguous, though, and modern scholarship is divided on whether Epicurus really made this claim and how it should be interpreted. I will critically review the relevant textual evidence as well as these modern interpretations and propose a compromise. Another epistemological problem concerns the claim, made by Diogenes of Oenoanda, that among alternative explanations some are more probable than others. I will argue that this claim finds no support in the writings of Epicurus and Lucretius and actually constitutes a departure from Epicurus’ views, for which I will suggest a possible motive. A related question concerns Bailey’s observation that Lucretius, in his astronomical passages, usually placed the view of the astronomers first, ‘as though he really preferred these.’ In a section devoted to this question I will propose a different interpretation, which will be more in line with Lucretius’ and Epicurus’ statements concerning the use of multiple explanations, and with Epicurus’ general attitude to mathematical astronomy.9 The remainder of the chapter is devoted to the question of Epicurus’ sources for the alternative explanations, firstly for the individual explanations, which are thought to have been largely 9

See esp. Sedley (1976).



drawn from doxography,10 and secondly for the systematic use of multiple alternative explanations, partial anticipations of which are suspected in Democritus, Aristotle and Theophrastus. In this context I will also deal with the ‘Syriac meteorology’, a meteorological treatise preserved in Syriac and Arabic, which consistently offers multiple alternative explanations. Although the manuscripts are unanimous in naming Theophrastus as its author, I think this ascription has been too readily accepted and the alternative hypothesis of an Epicurean origin, suggested by the earlier commentators, too readily rejected. Yet, the question of its authorship is important because the answer to this question largely determines our view as to the extent of Epicurus’ dependence on Theophrastus. A thorough comparison of the way multiple explanations are employed in the Syriac meteorology, in Epicurus’ Letter to Pythocles and Lucretius’ De rerum natura, and in the undisputed writings of Theophrastus, may bring us closer to an answer. Epicurus’ Letter to Pythocles and the sixth book of Lucretius’ De rerum natura are sometimes described as meteorological treatises. In this respect they can be compared with a number of other ancient texts, first and foremost Aristotle’s Meteorologica, but also the Syriac meteorology. Chapter two of my dissertation will be devoted to such a comparison, with a view to establishing the relations of these texts to one another. I will start with a thorough comparison of nine meteorological texts with regard to the range and subdivision of their subject matter, in order to distinguish the various traditions and the place of Epicurus and Lucretius therein. A characteristic feature of Lucretius’ meteorology is the attention he pays to exceptional local phenomena, which are either absent or far less conspicuous in other meteorologies and belong more properly to paradoxography.11 I will therefore continue the investigation with a comparison of the latter part of Lucretius’ book VI with a number of meteorological and paradoxographical works, with regard to the inclusion and treatment of exceptional local phenomena. Next I will deal with the order of meteorological subjects in Lucretius’ book VI, Aëtius’ book III, the Syriac meteorology, and Epicurus’ Letter to Pythocles. The large degree of correspondence between the first three works has often been observed, but never thoroughly examined; and the not-so-obvious resemblance of Epicurus’ Letter to the other three has generally been 10 11

Doxography = a genre of writings devoted to the collection of doxai or philosophical opinions: for a quick reference see Runia (1997b) or Mansfeld (2008). Paradoxography = a genre of writings devoted to the collection of paradoxa or marvellous stories: see Ziegler (1949); Schepens & Delcroix (1996), Wenskus & Daston (2000).



overlooked. I will therefore carry out a more thorough comparison of all four works with due attention to the similarities and differences in the order of their subjects, and try to establish an original order from which they all derive. This will be capped by an attempt to determine the precise relations between these four works and to identify possible missing links. In this context I will also continue my investigation into the authorship and identity of the Syriac meteorology by comparing its theories with, on the one hand, Epicurus and Lucretius, and, on the other hand, Theophrastus and Aristotle. In my third and final chapter I turn to one very specific cosmological problem: the shape of the earth. Although many scholars confidently claim that Epicurus and Lucretius believed the earth to be flat, and some of them even scorn them for having clung to such an antiquated idea, in reality no explicit statements about the shape of the earth can be found in their works. In this chapter I will consider this problem from various angles. I will re-examine the evidence that has been adduced so far for attributing a flat-earth cosmology to the Epicureans, and I will do so in the light of the two ancient cosmological systems distinguished by David Furley.12 In addition, I will discuss every passage in the works of Epicurus and Lucretius, as well as other Epicureans, that may be relevant to the question of the shape of the earth. I will also inventory the ancient proofs of the earth’s sphericity and search for clues that the Epicureans may have known, and responded to, these proofs. In the concluding section of this chapter I will bring all the evidence together and state my own opinion concerning the Epicurean view of the earth’s shape, a view which I shall also try to connect with the Epicureans’ general attitude towards astronomy. Although the three chapters of this dissertation constitute three separate studies, they are connected by several recurrent themes. One of these is the identity of the Syriac meteorology, which will be addressed in Chapters One and Two. Another is the Epicureans’ attitude towards mathematical astronomy, which will be addressed in a more general sense in Chapter One, while Chapter Three will deal with some particular instances. Yet another theme concerns the degree of Lucretius’ dependence on Epicurus: all three chapters touch upon passages in Lucretius that cannot be reduced to, or derived from, Epicurus: Lucretius’ passage on particular local phenomena in DRN VI 608ff (discussed in Chapters One and Two); his refutation of centripetal downward motion in DRN I 1052-93, and his cosmogonical account in DRN V 449-508 (both discussed in Chapter Three).


Furley (1986) and (1989a).

“I have devised seven separate explanations, each of which would cover the facts as far as we know them. But which of these is correct can only be determined by the fresh information which we shall no doubt find waiting for us.” - Sir Arthur Conan Doyle (1891-2), The Adventures of Sherlock Holmes: ‘The Adventure of the Copper Beeches’.

“That process,” said I, “starts upon the supposition that when you have eliminated all which is impossible, then whatever remains, however improbable, must be the truth. It may well be that several explanations remain, in which case one tries test after test until one or other of them has a convincing amount of support.” - Sir Arthur Conan Doyle (1921-7), The Case-Book of Sherlock Holmes: ‘The Adventure of the Blanched Soldier’.

1 MULTIPLE EXPLANATIONS 1.1 Introduction According to Epicurus and Lucretius, sunset and sunrise can be explained not only on the assumption that the sun passes unaltered below the earth and then emerges again, but equally well by its extinction and subsequent rekindling. Both options are retained because neither can be eliminated on the evidence of the senses, which is the Epicureans’ principal criterion of truth. This approach is typical of the way Epicurus, in the Letter to Pythocles, and Lucretius, in books V and VI of the De rerum natura, deal with astronomical, meteorological and terrestrial phenomena. Just like sunrise and sunset almost every one of these problems is accounted for with a number of alternative explanations, sometimes two, sometimes more. In defense of this method of multiple explanations Epicurus points out that in these fields of inquiry single explanations are neither possible nor necessary. They are not possible because the objects in question cannot be clearly observed because of their distance, and they are not necessary because the main aim of studying celestial and atmospherical phenomena is to rule out divine influence, for which it is enough to show that every phenomenon can be explained physically, although absolute certainty as to the actual cause is not needed. The method of multiple explanations depends on earlier philosophy in various ways. It has long been known that Epicurus and Lucretius derived the



vast majority of their alternative theories from the views of earlier philosophers, and there is good reason to believe that they learned of these views not from the original works but from doxographies, thematically organized collections of opinions of earlier thinkers. Nor were Epicurus and Lucretius the first to apply multiple explanations to certain problems. One notable instance is found among the fragments of Democritus and several occasional instances are found in the works of Aristotle and Theophrastus. The most complete parallel as to the use of multiple explanations is a meteorological treatise ascribed to Theophrastus and preserved in Syriac and Arabic, which employs multiple explanations on a similar scale as Epicurus and Lucretius, and which, if the ascription is correct, would make Theophrastus the real ‘inventor’ of the method of multiple explanations. In this chapter I propose to investigate the method of multiple explanations from several angles. First (in §1.2) I will briefly set out a number of variations in the use of multiple explanations. Then (in §1.3) I will deal at some length with a number of epistemological problems concerning the method, such as whether all alternative explanations are true, on what grounds theories are rejected or retained and what role analogy plays in all this, further whether the Epicureans allowed different degrees of probability, and whether in his astronomical passage Lucretius secretly preferred the theories of mathematical astronomy. In the middle part of the chapter (in §1.4) I will look into the relation of multiple explanations to doxography, and in the final part (§1.5) into the possible antecedents of the method of multiple explanations in Democritus, Aristotle and the undisputed writings of Theophrastus. In the same section I will also deal with the Syriac meteorology commonly ascribed to Theophrastus, and review the arguments for and against this ascription.

1.2 Variations in the use of multiple explanations Epicurus’ own methodological remarks on multiple explanations (Hdt. 7880 and Pyth. 2 [85-88]) seem to suggest that the method constitutes one unified and homogeneous whole. In fact, however, the lists of multiple explanations offered by Epicurus and Lucretius vary in several respects, the most important being number, exhaustivity, mutual exclusivity, subsidiarity and type-differentiation. Below I will briefly review these five kinds of variation and their possible significance for the present investigation. 1. Number. The total number of alternative explanations for a single problem varies from two to nine.13 For sunset and sunrise, for instance, both


See APPENDICES A & B on p.243 and p.245 below.



Epicurus and Lucretius offer two possible explanations.14 Thunder, on the other hand, is accounted for with no less than nine different explanations by Lucretius, though Epicurus gives only five.15 A few of the astronomical, meteorological and terrestrial problems discussed by Epicurus and Lucretius, e.g. the size of the heavenly bodies,16 the stability of the earth,17 and the temperature fluctuation in wells,18 are accounted for with only one explanation. In these cases we should distinguish between problems that may be open to other explanations and those which exclude them (see further below). 2. Exhaustivity. In most cases the lists of alternative explanations are not exhaustive, but appear to be open to other options as well, and Epicurus often tells us so explicitly.19 Only rarely do the alternative explanations seem to exhaust the entire range of possibilities, as when the moon is said to shine with either its own or with reflected light,20 or when the sun at sunset is said either to pass unaltered below the earth, or to be extinguished.21 This distinction of exhaustivity and inexhaustivity may perhaps be applied to single explanations as well (see above). Sometimes, although only one explanation is given, there is no need to suppose that this is the only one possible. This seems to be the case with a number of terrestrial phenomena discussed by Lucretius in the latter part of his book VI. Perhaps, because he was one of the first to discuss these problems, there was no anterior tradition from which he could derive his explanations (see p.113ff below), and he therefore contented himself with providing just one, and sometimes two. However, other instances of single explanations seem to exclude alternative views. Epicurus’ and Lucretius’ emphatic claim that the sun, the moon and the stars are the size they appear to be rules out all other options,22 and so does Epicurus’ account of the formation of the heavenly bodies out of windy and 14

Epic. Pyth. 7 [92]; Lucr. DRN V 650-679. Lucr. DRN VI 96-159: Ernout-Robin and Bailey distinguish only seven explanations, but admit that some of them are subdivided. I prefer to devide the account into nine separate explanations: (1) 96-115; (2) 116-120; (3) 121-131; (4) 132-136; (5) 137-141; (6) 142144; (7) 145-149; (8)150-155; (9) 156-159. Cf. Epic. Pyth. 18 [100]. 16 Epic. Pyth. 6 [91] and Lucr. DRN V 564-591. 17 Lucr. DRN V 534-563. 18 Lucr. DRN VI 840-7. 19 See APPENDIX A on p.243 below: a ‘+’-sign after the number indicates that the list is explicitly inexhaustive. 20 Epic. Pyth. 11 [94-95]. 21 See n.14 above. 22 See n.16 above. For a different interpretation of this theory see n.543 on p.214 below. 15



fiery matter.23 These two types of single explanations may perhaps be distinguished as accidentally single and necessarily single, only the latter being truly opposed to multiple explanations. 3. Mutual exclusivity. In some cases the alternatives offered seem to exclude each other: the sun is either extinguished at sunset or it is not, and the moon’s light is either borrowed or its own property.24 In most cases, however, nothing impedes the alternative explanations from obtaining at the same time. Once, in Pyth. 13 [96-97] on solar and lunar eclipses, Epicurus even tells us so explicitly: eclipses may be caused by extinction or interposition of another body, or both at the same time. 4. Subsidiarity. There are even some explanations that not merely allow, but actually require the simultaneous occurrence of another cause: a number of phenomena discussed in the latter part of DRN VI are accounted for with one principal explanation and one or two that are merely subsidiary to the first, and not capable of producing the desired effect on their own.25 Eruptions of the Etna, for instance, are said to be caused by wind blowing in caverns beneath the mountain, catching fire and then violently escaping, an effect that is further strengthened by incursions of the sea into these caverns.26 5. Type-differentiation. While in most cases any instance of a certain phenomenon can be accounted for by any one of the alternative explanations, sometimes – especially in meteorology – different explanations seem to apply to different types of the same phenomenon.27 Thunder, for instance, is explained according to the nature of its sound,28 lightning according to the presence or absence of thunder,29 and earthquakes according to their effects.30 Although Epicurus does not remark upon these differences, but presents multiple explanations as a single method, it seems useful to keep them in mind.

1.3 Truth, probability and personal preferences 1.3.1 Introduction Epicurus’ and Lucretius’ consistent use of multiple explanations in astronomy and meteorology is epistemologically very interesting if not 23

Epic. Pyth. 5 [90-91]. See n.14 and n.20 above. 25 See APPENDIX B on p.245 below. Cf. Bailey (1947) pp. 1655, 1684, 1704. 26 Lucr. DRN VI 639-702 27 Bailey (1947) 1567. 28 Lucr. DRN VI 96-159; cf. Bailey (1947) 1575. 29 Lucr. DRN VI 160-218; cf. Bailey (1947) 1578, 1586. 30 Lucr. DRN VI 535-607. 24



problematic. There is, for instance, a continuing debate on whether Epicurus and Lucretius considered all alternative explanations merely possible or actually true, and, if true, how these simultaneous truths should be conceived of, and on what grounds Epicurus felt he could pronounce some explanations true and others false. Conversely, if alternative explanations are only possible, are they all equally possible or do they admit different degrees of probability, and is it permitted for an Epicurean to prefer one explanation over another? These are some of the problems I propose to deal with below. The section will be structured as follows. First (in §1.3.2) I will discuss the question whether Epicurus considered all alternative explanations true. Then I will deal with the grounds for rejecting theories (in §1.3.3) and the role played by analogy in accepting theories (in §1.3.4). Following this (in §1.3.5) I will consider whether or not the Epicureans allowed different degrees of probabilities and personal preferences, and having answered this in the negative I will (in §1.3.6) critically examine Bailey’s claim that in astronomy Lucretius had a secret preference for the views of the mathematical astronomers.

1.3.2 Are all alternative explanations true? It is often claimed that Epicurus considered all alternative explanations true. Although we do not have any explicit statement by Epicurus to this effect, it seems to follow logically from his use of non-contestation to confirm individual alternative theories and his claim that non-contestation establishes truth.31 On the other hand, the simultaneous truth of several, often mutually exclusive, explanations seems to violate the principle of non-contradiction, to which Epicurus was also committed.32 In this subsection I propose to examine whether or not Epicurus might have claimed that all alternative explanations are true. In order to do so I will first (in review the relevant ancient texts, then (in present and assess the most important modern theories and formulate a provisional conclusion, then (in voice some reservations about this conclusion, and finally (in present a more definitive conclusion.


On non-contestation and multiple explanations in Epicureanism see e.g. Striker (1974) and (1996); Sedley (1982) 263-72; Long & Sedley (1987), vol. I, 90-97; Asmis (1984) 178-80, 193-96, 211, 321-36; id. (1999) 285-94; and Allen (2001) 194-205 & 239-41; Fowler (2002) 191-92. 32 See Asmis (1984) 194. See also n.35 below.


13 Ancient texts and testimonies The various modern views concerning the truth-value of multiple explanations are based on a number of ancient texts and testimonies, the most important of which will be presented below. A. The principles of Epicurean epistemology According to our sources,33 Epicurus acknowledged three criteria of truth: perceptions (afisyÆseiw), preconceptions (prolÆceiw) and affections (pãyh). Perceptions are the raw data presented to us by the senses; in addition Epicurus distinguished a kind of ‘mental perception’, called fantastikØ §pibolØ t∞w diano¤aw (‘impressional projection of the mind’), which directly affects the mind, without the mediation of the senses, such as we experience in dreams; later Epicureans made this into a separate criterion of truth. Preconceptions are general notions naturally formed in our minds in response to repeated perceptions, and affections are the primary emotions – pleasure and pain – by which choice and avoidance are determined. These three, or four, types of data, which are themselves incontrovertibly true, serve as the criteria by which the truth and falsity of opinions are established. For the investigation of physical reality not all criteria are equally relevant. Affections are mainly of use in ethics. ‘Mental perceptions’ are useful insofar as they help us to form a preconception of the gods, including the realisation that the gods are in no way responsible for the occurrences in our world. Preconceptions provide us with the necessary notions to be able to formulate opinions at all: we can only investigate the causes of thunder if we have a clear notion of what thunder is. However, supposing that the necessary preconceptions are there, and that hypotheses concerning physical reality can be formulated, the criteria by which these are tested are perceptions and ‘things perceived’, i.e. appearances (fainÒmena).34 Opinions are either true or false.35 An opinion is true when it is attested (§pimartur∞tai) or not contested (mØ éntimartur∞tai) by appearances, and false when it is not attested (mØ §pimartur∞tai) or when it is contested (éntimartur∞tai) by appearances.36 Opinions to be tested fall into one of two 33

The most important sources for Epicurus’ epistemology are Diog. Laërt. X 31-34 and Sext. Emp. Math. VII 203-211, supplemented at crucial points by quotations from Epicurus himself. 34 I translate fainÒmena as ‘appearances’, which captures the general sense better than ‘phenomena’. 35 Diog. Laërt. X 34; Sext. Emp. Math. VII 211. See, however, Cic. De fato 19, 21, 37; Acad. pr. II 97; De nat. deor. I 70 (collected as Epic. fr.376 Us.), where it is stated that Epicurus refused to assign any truth value to opinions about the future. 36 Epic. Hdt. 51: §ån m¢n mØ §pimarturhyª µ éntimarturhyª, tÚ ceËdow g¤netai: §ån d¢ §pimarturhyª µ mØ éntimarturª, tÚ élhy°w. – “if it is not attested, or is contested,



categories: the prosm°nonta, those ‘awaiting’ confirmation by a closer and clearer observation, and the êdhla, the ‘unclear’ or ‘non-apparent’, which do not allow closer observation.37 The latter can be further subdivided into those which can only be observed from afar, such as astronomical and meteorological phenomena (met°vra),38 and those which cannot be observed at all, such as the existence of atoms and void, and other theories fundamental to Epicurean physics.39 Of the two categories mentioned above, the prosm°nonta are typically tested by attestation and non-attestation,40 the êdhla by non-contestation and contestation.41 B. Sextus Empiricus on Epicurus’ conditions of truth and falsehood The only complete account of Epicurus’ four methods of verification and falsification that has come down to us is provided by Sextus Empiricus in Math. VII 211-216.42 According to Sextus, attestation occurs when a hypothesis about something is confirmed by a closer observation of the object in question, e.g. when we see someone approaching from afar and hypothesize that it is Plato, this hypothesis is confirmed when he has come closer and is seen to be really Plato.43 When, on the other hand, on approaching he turns out not to be Plato, the hypothesis is rejected by non-attestation.44 Note that Sextus’ use of the term non-attestation is more restrictive than the words themselves suggest: non-attestation seems to denote not merely the negation or absence of attestation but rather the attestation of the negated hypothesis. Of contestation Sextus gives the following account (214):

falsehood arises; but if it is attested or not contested, truth.” Cf. Diog. Laërt. X 34; Sext. Emp. Math. VII 211 & 216. The translations of the technical terms are those of Sedley (1982), Long & Sedley (1987) and Allen (2001). Asmis (1984) & (1999) prefers witnessing, no-counterwitnessing, no-witnessing, counterwitnessing. 37 Epic. Hdt. 38; id. RS 24. Cf. Diog. Laërt. X 34. 38 Epic. Hdt. 80 (... afitiologht°on Íp°r te t«n mete≈rvn ka‹ pantÚw toË édÆlou ...). 39 Epic. Hdt. 38, introducing the fundamental theories of Hdt. 38-44. 40 Epic. RS 24. Cf. Diog. Laërt. X 34. 41 Epic. Pyth. 2 [88], 3 [88], 7 [92] (non-contestation applied to astronomical and meteorological phenomena); Sext. Emp. Math. VII 213-14 (contestation and noncontestation applied to what is by nature unobservable). 42 = Long & Sedley 18A = Usener 247 (part). 43 Sext. Emp. Math. VII 212. 44 Sext. Emp. Math. VII 215.

MULTIPLE EXPLANATIONS ÑH m°ntoi éntimartÊrhsiw {...} ∑n går énaskeuØ toË fainom°nou t“ Ípostay°nti édÆlƒ, oÂon ı StvikÚw l°gei mØ e‰nai kenÒn, êdhlÒn ti éji«n, toÊtƒ d¢ oÏtvw Ípostay°nti Ùfe¤lei tÚ fainÒmenon sunanaskeuãzesyai (fhm‹ d' ≤ k¤nhsiw): mØ ˆntow går kenoË kat' énãgkhn oÈd¢ k¤nhsiw g¤gnetai.


Contestation {...} was the elimination of the appearance with the hypothesized non-evident fact, as for instance, when the Stoic says there is no void, claiming something non-evident, the appearance (I mean motion) must be co-eliminated with what is thus hypothesised, for if there is no void, by necessity motion doesn’t occur either.

A hypothesis about something non-evident (êdhlon) is proved wrong by contestation when its acceptance would lead to the elimination or cancellation of an evident fact. The argument can be set out as follows: ¬v → ¬m m ———— v

if there were no void, there would be no motion but there is motion ———————————————————— therefore there is void

If the necessity of the first premise, and the evidence of the second are accepted, the conclusion must be true. Non-contestation is explained by Sextus as follows (213-4): OÈk éntimartÊrhsiw d° §stin ékolouy¤a toË Ípostay°ntow ka‹ dojasy°ntow édÆlou t“ fainom°nƒ, oÂon ı ÉEp¤kourow l°gvn e‰nai kenÒn, ˜per §st‹n êdhlon, pistoËtai di' §nargoËw prãgmatow toËto, t∞w kinÆsevw: mØ ˆntow går kenoË oÈd¢ k¤nhsiw ˇˇ≈feilen e‰nai, tÒpon mØ ¶xontow toË kinoum°nou s≈matow efiw ˘n peristÆsetai diå tÚ pãnta e‰nai plÆrh ka‹ nastã, Àste t“ dojasy°nti édÆlƒ mØ éntimarture›n tÚ fainÒmenon kinÆsevw oÎshw.

Non-contestation is the attendance of the hypothesized and supposed non-evident fact upon the appearance, as for instance, when Epicurus says there is void, which is non-evident, this is proved by an evident thing, motion: for if there were no void, there shouldn’t be motion either, as the moving body wouldn’t have a place into which to come round, because everything would be full and packed; therefore the appearance does not contest the supposed non-evident thing, since there is motion.

A hypothesis about something non-evident (êdhlon) is proved right by non-contestation when it can be shown to ‘attend upon’, or be implied by, an evident appearance, as, for instance, the non-evident existence of void follows from the evident existence of motion. Sextus’ general account of noncontestation suggests the following schema (modus ponens): m→v m ——— v

if there is motion, there is void there is motion ————————————— therefore there is void

This schema, however, fails to make clear the function of the negation in non-contestation. If we follow Sextus’ example rather than his theoretical



account, we find a different, though logically equivalent, schema (modus tollens): ¬v → ¬m m ———— v

if there were no void, there would be no motion but there is motion ———————————————————— therefore there is void

This schema is exactly identical to that of contestation. The only difference is one of focus: contestation is about falsifying a hypothesis (viz. the nonexistence of void), whereas non-contestation is about verifying a hypothesis (viz. the existence of void), by falsifying its negation (viz. the non-existence of void). In other words, according to Sextus’ account, non-contestation does not just denote – as the name suggests – the absence of contestation, but contestation of the negated hypothesis. C. Epicurus’ account of the fundamental theories of physics In Hdt. 38-44, Epicurus discusses, under the general heading of êdhla, i.e. non-evident things, the principal and fundamental tenets of his physical theory (e.g. that nothing comes from nothing, that the sum total of things is unchanging, that everything consists of bodies and void, etc.). One of the subjects discussed is the existence of void, which Epicurus sets out as follows (Hdt. 40)45: efi d¢ mØ ∑n ˘ kenÚn ka‹ x≈ran ka‹ énaf∞ fÊsin Ùnomãzomen, oÈk ín e‰xe tå s≈mata ˜pou ∑n oÈd¢ diÉ o §kine›to, kayãper fa¤netai kinoÊmena.

and if there were not what we call void and space and intangible nature, bodies would have no place to be or through which to move, as they are observed to move.

Except for the suppressed conclusion, viz. that void exists, this argument is identical, both in subject and structure, to Sextus’ example of non-contestation (item B on p.14 above), and most of Epicurus’ other arguments in this section exhibit the same structure. Epicurus does not, however, in this context speak of non-contestation, nor does he provide any other name for the procedure. A little further on Epicurus discusses the sizes of atoms (Hdt. 55-56):


Cf. Lucr. DRN I 334-45. On this argument see Allen (2001) 195-6; Furley (1971), and Furley’s response to Schrijvers in Gigon (1978), 117-8.

MULTIPLE EXPLANATIONS ÉAllå mØn oÈd¢ de› nom¤zein pçn m°geyow §n ta›w étÒmoiw Ípãrxein, ·na mØ tå fainÒmena éntimarturª, {...} pçn d¢ m°geyow Ípãrxon oÎte xrÆsimÒn §sti prÚw tåw t«n poiotÆtvn diaforãw, éf›xya¤ te ëmÉ ¶dei ka‹ prÚw ≤mçw ıratåw étÒmouw, ˘ oÈ yevre›tai ginÒmenon, ... .


Nor, moreover, must we suppose that every size exists among the atoms, lest the appearances contest this, {...} but the existence of every size of atoms is not required for the differences of their qualities, and at the same time visible atoms would have to come within our ken, which is not observed to happen, ...

Here Epicurus has interwoven two different arguments. In the first place the existence of atoms of every size is unnecessary for his physical theory, and in the second place it would entail the existence of observable atoms,46 which is in conflict with the evidence of the senses. This second part of the argument can be set out schematically as follows: p→q ¬q ——— ¬p

if atoms could have every size, some atoms would be observable but there are no observable atoms ———————————————————————— therefore atoms cannot have every size

The argument closely matches Sextus’ account of contestation, and – what is more – this time Epicurus himself refers to the argument by this very term: mØ tå fainÒmena éntimarturª – ‘lest the appearances contest this’. May we then conclude that Sextus has correctly reported Epicurus’ views on contestation and non-contestation? D. Epicurus’ use of non-contestation in astrononomy and meteorology Until now contestation and non-contestation have only been applied to fundamental physical theories. It remains to be seen how they are used in astronomy and meteorology. In Pyth. 7 [92], Epicurus offers two alternative explanations to account for the risings and settings of the sun, the moon and the stars:


Epicurus has ‘visible atoms’, but on his own theory (cf. Hdt. 46-50 and Lucr. IV 54-216) even enormous atoms would not be directly ‘visible’, since being atomic they would not be able to shed the necessary images, but they would be indirectly ‘visible’, by blocking other things from view, and also tangible because of all the senses touch alone does not require the shedding of particles.



ÉAnatolåw ka‹ dÊseiw ≤l¤ou ka‹ selÆnhw ka‹ t«n loip«n êstrvn (1) ka‹ katå ênacin g¤nesyai dÊnasyai ka‹ katå sb°sin, toiaÊthw oÎshw peristãsevw ka‹ kay' •kat°rouw toÁw tÒpouw, Àste tå proeirhm°na épotele›syai: oÈd¢n går t«n fainom°nvn éntimarture›: (2) kat' §kfãneiãn te Íp¢r g∞w ka‹ pãlin §piprosy°thsin tÚ proeirhm°non dÊnait' ín suntele›syai: oÈd¢ gãr ti t«n fainom°nvn éntimarture›.

Risings and settings of the sun and the moon and the other heavenly bodies (1) may come about by kindling and extinction, the circumstances at both places [i.e. the places of rising and setting] being such as to produce the afore-mentioned results: for nothing in appearances contests this. (2) and by their appearance above the earth and again the (earth’s) interposition the aforementioned result might be produced: for not a thing in appearances contests this.

According to Epicurus, the heavenly bodies are either repeatedly extinguished and then rekindled, or they pass unaltered below the earth and then emerge again.47 Both options must be accepted, because ‘nothing in appearances contests’ either one of them. The same or similar terms, expressing either the absence of disagreement48 or the presence of (positive) agreement with appearance,49 occur throughout the Letter to Pythocles, often to account for each one of a number of alternative explanations. If, as Epicurus claims,50 non-contestation establishes truth, then each one of the alternative explanations must be true. It should be noted, however, that Epicurus’ argument in these passages is not at all like Sextus’ account of noncontestation (item B on p.14 above); nowhere in the Letter to Pythocles do we encounter anything that resembles the syllogistic structure of non-contestation as set out by Sextus and as applied (though without this ‘label’ of noncontestation) by Epicurus in the Letter to Herodotus to prove his fundamental physical theories. In the present context non-contestation seems to be nothing more than the absence of contestation by appearances, or – in other words – the (positive) agreement with appearances. E. Epicurus on the distinction between single and multiple explanations It would seem then, that Epicurus dealt differently with the fundamental physical theories and the more specialised theories concerning astronomical 47

Similarly Lucr. DRN V 650-679 on the causes of nightfall and dawn. Pyth. 2 [88] « oÈk éntimarture›tai », Pyth. 3 [88] « t«n går fainom°nvn oÈd¢n éntimarture› », Pyth. 9 [93] « oÈyen‹ t«n §narghmãtvn diafvne› », Pyth. 11 [95] « oÈy¢n §mpodostate› t«n §n to›w mete≈roiw fainom°nvn », Pyth. 16 [98] « oÈ mãxetai to›w fainom°noiw ». 49 Pyth. 2 [86] « to›w fainom°noiw sumfvn¤an »; Pyth. 2 [87] « sumf≈nvw to›w fainom°noiw » and « sÊmfvnon ... t“ fainom°nƒ »; Pyth. 9 [93], 12 [95], 32 [112] « tÚ sÊmfvnon to›w fainom°noiw ». 50 See n.36 on p.13 above. 48



and meteorological phenomena. Whereas the first are proved to be uniquely true by showing that their negation is contested by appearances, the second are accounted for with a number of alternative theories which must all be accepted because none of them is contested by appearances. Epicurus explicitly contrasts the two types of problems in Pyth. 2 [86]: MÆte tÚ édÊnaton [ka‹] parabiãzesyai, mÆte ımo¤an katå pãnta tØn yevr¤an ¶xein µ to›w per‹ b¤vn lÒgoiw µ to›w katå tØn t«n êllvn fusik«n problhmãtvn kãyarsin, oÂon ˜ti tÚ pçn s«ma ka‹ énafØw fÊsiw §st¤n, µ ˜ti êtoma stoixe›a ka‹ pãnta tå toiaËta dØ ˜sa monaxØn ¶xei to›w fainom°noiw sumfvn¤an: ˜per §p‹ t«n mete≈rvn oÈx Ípãrxei, éllå taËtã ge pleonaxØn ¶xei ka‹ t∞w gen°sevw afit¤an ka‹ t∞w oÈs¤aw ta›w afisyÆsesi sÊmfvnon kathgor¤an.

We must not try to force an impossible explanation, nor employ a method of inquiry similar in every respect to our reasoning either about the modes of life or with respect to the sorting-out of other physical problems, such as our statement that ‘the universe consists of bodies and the intangible’, or that ‘the elements are indivisible’, and all such statements which exhibit a singular agreement with appearances. For this is not so with the things above us: they admit of a plural account of their cominginto-being and a plural expression of their nature which agrees with our sensations.

Althought the fundamental physical theories (just like the theories concerning the modes of live) no less than astronomical and meteorological theories need to fulfill the requirement of agreement with appearances, the method of inquiry by which they are approached is different. In other words: Epicurus explicitly recognizes the existence of two different methods of inquiry, one that is applied to such problems as admit only one answer (such as the fundamental physical theories), and one that is applied to those problems that admit several answers (such as we find in astronomy and meteorology). F. Lucretius on the truth of all alternative explanations (1) Lucretius too offers multiple explanations to account for astronomical (DRN V 509-770) and meteorological (DRN VI) phenomena. Close to the beginning of his astronomical section, having just offered a number of alternative explanations for the movements of the stars, he states his view on the epistemological status of these explanations (V 526-33): nam quid in hoc mundo sit eorum ponere certum difficilest; sed quid possit fiatque per omne in variis mundis varia ratione creatis, id doceo plurisque sequor disponere causas, motibus astrorum quae possint esse per omne; 530 e quibus una tamen siet hic quoque causa necessest, quae vegeat motum signis; sed quae sit earum praecipere haudquaquamst pedetemptim progredientis.

for to state with certainty which of these causes holds in our world is difficult; but what can and does happen throughout the universe in the various worlds created in various ways, this I teach, and I proceed to set forth several causes for the motions of the stars, which may apply throughout the universe; one cause out of this number, however, is necessarily the case here too, which gives force to the motion of the stars, but which of them it is, is not for them to lay down who proceed step by step.

Although in our world each explanation can at best be called possible, in the universe at large, given the infinity of space and matter and hence of



worlds, any given possibility cannot fail to be realised (the ‘principle of plenitude’),51 and so every possible explanation is also ‘true’, if not here, then somewhere else. This may also explain why Epicurus in his Letter to Pythocles sometimes speaks of the alternative explanations as all being actually the case and coexisting rather than being merely possible and mutually exclusive.52 G. Lucretius on the truth of all alternative explanations (2) There is a second passage where Lucretius deals with the method of multiple explanations. In VI 703-711, preceding his account of the summer flooding of the Nile, Lucretius writes: Sunt aliquot quoque res quarum unam dicere causam non satis est, verum pluris, unde una tamen sit; corpus ut exanimum siquod procul ipse iacere conspicias hominis, fit ut omnis dicere causas conveniat leti, dicatur ut illius una; nam que eum ferro nec frigore vincere possis interiisse neque a morbo neque forte veneno, verum aliquid genere esse ex hoc quod contigit ei 710 scimus. item in multis hoc rebus dicere habemus.

There are also a number of cases for which naming one cause is not enough, but several, one of which is nevertheless the case; just as, if you should yourself see a person’s dead body lying at a distance, it happens to be fitting to name all the causes of death, to make sure that the one cause of this death be named; for you could not prove that he died from the sword or from cold or from disease or perchance from poison, but we know that it was something of this sort that befell him. Similarly we must say this in many cases.

Although this passages is not incompatible with the earlier one, there is an interesting shift of focus. This time we hear nothing of the infinite number of worlds, nothing of the principle of plenitude, and nothing of the truth of all explanations, and, although the event is still viewed as a particular instant of a certain class of events, the emphasis seems to be on the particular instance, to which only one explanation applies (although we do not know which one), rather than the whole class of events, to which many explanations apply. Moreover, the chosen example, a dead body, seems strangely inappropriate. While multiple explanations are typically applied to things that can only be 51

Cf. Lucr. DRN V 422-31 and Epic. fr.266 Us. (both referring to the infinity of time rather than matter and space). On the Epicurean use of the principle of plenitude see e.g. Sedley (1998a) 175, n.29. A precursor of this principle (based on the infinity of space and matter) is described by Aristotle (Phys. III, 4,203b25-30), who may be rendering a Democritean view: see Asmis (1984) 264-5. 52 In the Letter to Pythocles I have counted 34 cases where multiple explanations are offered. 20 of these are accompanied by some verb or expression denoting possibility (§nd°xetai, dÊnatai, oÈk édÊnaton, etc.). Of the 14 remaining cases, 5 exhibit a purely disjunctive list of alternative explanations (≥toi A µ B µ G). The 9 remaining cases are either conjunctive (ka‹ A ka‹ B ka‹ G) or mixed (ka‹ A ka‹ B µ G): in these 9 cases the language seems to suggests that several explanations may be true at the same time. For more details see APPENDIX A on p.243 below.



seen from a distance, there does not seem to be any cogent reason why the dead body could not be approached and examined more closely, so as to eliminate certain explanations and perhaps even arrive at the one true cause of death (concerning this example see also §, fourth paragraph, below). Three different modern theories The main problem with the above texts and passages concerns items B and D (p.14 and p.17 respectively): Sextus’ account of non-contestation (B) and Epicurus’ use of non-contestestion in the Letter to Pythocles (D) seem to be incompatible. According to Sextus, non-contestation is contestation of the negated hypothesis, which by means of a syllogism establishes the exclusive truth of its hypothesis, thus satisfying Epicurus’ claim (item A on p.13 above) that non-contestation establishes truth. In the Letter to Pythocles, on the other hand, non-contestation seems to mean nothing more than the absence of contestation, which – common-sense suggests – can only establish the possibility of each of a number of alternative explanations. Yet, Lucretius (item F on p.19 above) suggests a way in which possibility may actually amount to truth, to the effect that even non-contestation in this weak sense might satisfy Epicurus’ claim (item A) about non-contestation being a condition of truth. Several ways to resolve the observed incompatibility can be and have been devised: 1. Sextus’ account of non-contestation is correct, and Epicurus’ use of non-contestation in the Letter to Pythocles must be interpreted accordingly: non-contestation does not apply to each alternative explanation in isolation, but to the entire disjunction of alternative explanations. 2. Non-contestation as explained by Sextus and non-contestation as applied in Epicurus’ Letter to Pythocles are two different things: Sextus’ account is about non-contestation in the strong sense which establishes the truth of a proposition, whereas Epicurus’ Letter to Pythocles refers to non-contestation in the weak sense which only establishes possibility. 3. Sextus’ account of non-contestation is incorrect or at least incomplete. Epicurus’ own use of non-contestation in the Letter to Pythocles is our only certain guide to the working of non-contestation as perceived by Epicurus. Below I will examine each of these three approaches more thoroughly. Ad 1. An interesting solution to the problem is offered by Jim Hankinson.53 According to Hankinson, an Epicurean explanation of an atmospherical or 53

Hankinson (1999a) 221-23, and (1999b) 505-7.



celestial phenomenon takes the form of a disjunction of possible explanations: “x occurs because either E1 or E2 or ... En. At most one of the Ei’s can be the true explanation (cf. Lucretius 6. 703-4); but if the disjunction is sufficiently all-embracing, one of them will be: and that is all that is required.” Tad Brennan,54 elaborating on Hankinson’s remark, adds: “the point could be strengthened by reflecting that the actual reference in DL 10.86 [Pyth. 2] does not mention multiple “aitiai”, plural, but a “pleonachên aitian”,55 i.e. a single explanation with a complex, manifold structure. This is why the assertion of the whole disjunction is safe but the isolated assertion of any one disjunct is not (DL 10.87) [Pyth. 2].” In other words, if complete, the whole disjunction, i.e. the entire range of possible explanations, can be called true, and could in principle be demonstrated to be so by non-contestation (as interpreted by Sextus Empiricus). In this context Brennan might also have quoted Pyth. 2 [88], where non-contestation seems to be applied to the fact of there being multiple explanations, not to any single explanation in particular: TÚ m°ntoi fãntasma •kãstou thrht°on ka‹ §p‹ tå sunaptÒmena toÊtƒ diairet°on, ì oÈk éntimarture›tai to›w par' ≤m›n ginom°noiw pleonax«w suntele›syai.

Yet the appearance of each appearance must be preserved, and, as regards what is associated with it, those things must be distinguished whose production in a multiple way is not contested by phenomena here with us.

Hankinson’s account, even when supplemented by Brennan’s remarks, is disappointingly short. It would be interesting to see their interpretation of noncontestation applied to a specific case in the Letter to Pythocles. Unfortunately it is very hard to find a case where the disjunction of possible explanations logically follows from an evident fact, as Sextus’ account of non-contestation prescribes. It is at this point that we may call another text to our aid. Among the many papyri which have been unearthed at Herculaneum, there is one which preserves part of a treatise by Philodemus on the Epicurean theory of signs.56 In one of the fragments two different types of inference are distinguished.57 The first type, which is called method of elimination (ı katå tØn énaskeuØn trÒpow), is illustrated by the following example: ‘if there is motion, there is void’. This is the type which, we have seen, underlies Sextus’ accounts of contestation and non-contestation. Beside this, there is another type, which is called method of similarity (ı katå tØn ımoiÒthta trÒpow), 54

Brennan (2000) commenting on Hankinson (1999b) 505-7. See the text quoted in item E on p.18 above. 56 Philodemus De signis: complete edition with English translation in De Lacy (1978). 57 De signis xi 32 – xii 31 (= Long & Sedley 18F) et passim. 55



which is illustrated by the following example: ‘if Plato is a human being, Socrates is a human being’. Although Socrates’ humanity does not strictly follow from Plato’s, it is inconceivable that the one should be human and the other not. Another, more general example of this type of inference is: ‘if men here with us are mortal, then men everywhere are mortal’.58 Philodemus argues against certain opponents that both types are equally valid as methods of inference. For the present purpose I will assume that not only Philodemus’ method of elimination, but also his method of similarity can provide a valid basis for non-contestation as described by Sextus.59 It is now time to return to the promised example of non-contestation being applied to the whole disjunction of possible explanations. In Pyth. 11 [94-5] Epicurus writes about the light of the moon: ÖEti te §nd°xetai tØn selÆnhn §j •aut∞w ¶xein tÚ f«w, §nd°xetai d¢ épÚ toË ≤l¤ou. ka‹ går par' ≤m›n yevre›tai pollå m¢n §j •aut«n ¶xonta, pollå d¢ éf' •t°rvn.

Next, the moon may have her light from herself or may have it from the sun. For here with us, too, we see many things having light from themselves, and many having it from something else.

If we accept that no other explanations but these two are possible, we can formulate the following disjunction: ‘the moon has its light either from itself or from something else’. On the basis of this disjunction we can now construct the following implication (of the similarity type): ‘if light-giving objects here with us do so either because they have light from themselves or from something else, then the moon (being a light-giving object) does so either because it has light from itself or from something else’. If we accept the implication, then, as the antecedens is evident, the consequens must be true as well. The truth of the disjuntion has been verified by non-contestation (according to Sextus). This can be set out schematically as follows (modus ponens): p → (q1 ∨ q2) p ————— (q1 ∨ q2)

if the moon gives light (p), it has this light either from itself (q1) or from something else (q2) the moon gives light (p) ————————————————————————————— the moon has its light either from itself (q1) or from something else (q2)

This procedure can also be applied to most other cases where multiple explanations are in order, not only those with a limited number of alternatives, but also the inexhaustive disjunctions, where Epicurus explicitly tells us that still other explanations can be added.60 58

De signis xvi 5-29 et passim. See Asmis (1984) 202. 60 See APPENDIX A on p.243 ff below. 59



However, there is one major problem to this interpretation of noncontestation as applied to multiple explanations: it does not take into account the fact that in the Letter to Pythocles non-contestation is sometimes invoked to prove each one of a number of alternative explanations rather than the disjunction as a whole (cf. item D on p.17 above). Ad 2. Another approach, proposed by Gisela Striker and followed by Don Fowler,61 is to simply acknowledge the existence, in Epicurean epistemology, of two different kinds of non-contestation. On the one hand there is noncontestation in the strong sense, meaning contestation of the negated hypothesis, which is described by Sextus Empiricus (item B on p.14 above) and repeatedly applied (though not by name) by Epicurus himself in the Letter to Herodotus (item C on p.16 above). It is with reference to this type of noncontestation that Epicurus calls non-contestation a condition of truth (item A on p.13 above). On the other hand there is non-contestation in the weak sense, which is simply the negation or absence of contestation, and which is invoked several times in Epicurus’ Letter to Pythocles to establish the possibility of each of a number of alternative explanations of astronomical and meteorological phenomena (item D on p.17 above). Only with reference to the ‘principle of plenitude’ (item F on p.19 above) can all these alternative explanations also be called true, if not in this world, then in another.62 This approach has the great advantage of preserving all available evidence: it allows us to accept Sextus’ account of non-contestation and at the same time do justice to Epicurus’ own use of non-contestation in the Letter to Pythocles. The great drawback is that it leaves us with two different kinds of noncontestation not distinguished by name, but having quite different logical outcomes, despite Epicurus’ unqualified claim that non-contestation establishes truth. Ad 3. A third approach is to simply dismiss Sextus’ account of noncontestation as incorrect or inaccurate. This has been, with minor variations, the approach of David Sedley, Elizabeth Asmis and James Allen.63 According to this approach, there is only one kind of non-contestation, viz. the method which Epicurus himself employs several times in his Letter to Pythocles, where non-contestation amounts to nothing more specific than agreement or 61

Striker (1974) 76; id. (1996) 45; Fowler (2002) 191-92. Striker (1974) 78-9; id. (1996) 47-8. 63 Sedley (1982) 263-72; Long & Sedley (1987), vol. I, 90-97; Asmis (1984) 178-80, 19396, 211, 321-36; id. (1999) 285-94; and Allen (2001) 194-205 & 239-41. 62



compatibility with appearances and may result in the acceptance of several theories at the same time. It is to this procedure that Epicurus’ claim that noncontestation establishes truth must be applied. It cannot be denied that the method which Sextus Empiricus describes under the heading of ‘non-contestation’ corresponds very well to the way Epicurus, in the Letter to Herodotus, proves many of his fundamental physical theories by showing that their contradictories are contested by the appearances (item C on p.16 above). It is also true, however, that in the Letter to Pythocles Epicurus makes the fundamental physical tenets, no less than astronomical and meteorological theories, subject to agreement with appearances (item E on p.18 above). This apparent contradiction can be explained, basically, in two ways: either (a) Epicurus did not consider contestation of the contradictory a real proof, despite his repeated use of this method in the Letter to Herodotus, believing that the real proof consisted only in non-contestation, i.e. agreement with appearances, which – curiously – he often fails to invoke in the Letter to Herodotus, or (b) he though that contestation of the contradictory by itself somehow implied, and therefore could be subsumed under, non-contestation. The first option is chosen by Sedley, who tries to minimize the importance of contestation of the contradictory for Epicurus, pointing out that the logical implication on which this procedure rests, e.g. ‘if there is no void, there is no motion’, must itself be proved by non-contestation, i.e. agreement with appearances,64 and therefore cannot count as a condition of truth. Against this position I would like to stress the following two points: firstly that Epicurus in the Letter to Herodotus (see item C on p.16 above) repeatedly uses contestation of the contradictory as a proof in its own right, presenting the underlying implications as self-evident, rather than requiring further proof; and secondly that Epicurus in the Letter to Pythocles (2 [86]: see item E on p.18 above) explicitly distinguishes two methods of inquiry, one applied to the fundamental physical problems and resulting in single explanations, the other to meteorological and astronomical problems and resulting in multiple explanations. The second option, or something like it, is advocated by Asmis and Allen. Asmis acknowledges the importance of contestation of the contradictory, but generally refers to it as a kind of (positive) contestation (rather than noncontestation as Sextus does) which she opposes to the method of ‘induction’ (i.e. agreement with appearances, or non-contestation).65 In a later publication, however, Asmis seems to subsume both types of scientific inference under the general heading of non-contestation, which makes her position come very 64 65

Sedley (1982) 269 with n.70; cf. Allen (2001) 203. Asmis (1984) 211 et passim.



close to Gisela Striker’s (see Ad 2 above).66 Allen suggests that for Epicurus proofs by contestation of the contradictory would have been a special case of the more common proof by agreement with appearances.67 Asmis’ and Allen’s view may perhaps be summarized as follows: Epicurus claims that in matters of the non-evident truth is established by non-contestation, by which he means agreement with appearances. This agreement with appearances may be singular, as with the fundamental physical theories, or plural, as with most astronomical and meteorological phenomena. To establish a theory’s singular agreement with appearances Epicurus uses a method that may be described as contestation of the contradictory hypothesis, which he seems to have considered a special kind of non-contestation. Sextus’ report would then be inaccurate insofar as it identifies this contestation of the contradictory hypothesis with non-contestation, of which it is only a special kind. Having reviewed a number of modern views concerning the truth of multiple explanations it is now time to summarize the results. The first approach, defended by Jim Hankinson and Tad Brennan, claiming that noncontestation does not apply to, and therefore does not establish the truth of, individual alternative explanations, but applies only to the complete disjunction of alternatives, is refuted by the evidence. A second approach, defended by Gisela Striker, assuming two different kinds of non-contestation, one of which establishes the truth of single theories, while the other only establishes the possibility of each one of a number of alternative explanations, introduces a distinction that seems unwarranted by Epicurus’ unqualified claim that non-contestation establishes truth.68 This leaves us only the third approach, which maintains that Epicurus really claimed the truth of all alternative explanations. As we saw above two varieties of this approach can be distinguished. The first variety, defended by David Sedley, rejects not only Sextus’ account of non-contestation as contestation of the contradictory hypothesis, but also minimizes the importance of Epicurus’ frequent use of contestation of the contradictory hypothesis to prove the most fundamental physical theories. At the same time it seems to ignore Epicurus’ explicit distinction of two methods of inquiry, instead suggesting that truth is utimately always established by agreement with appearances. The second variety, which may be attributed to Elizabeth Asmis and James Allen, deals more cautiously with the available evidence. It provides a plausible way in which contestation 66

Asmis (1999) 289. Allen (2001) 200. 68 See n.36 on p.13 above. 67



of the contradictory hypothesis may at once be separate from, and yet subsumed under, non-contestation. On this interpretation Sextus’ report of non-contestation can be retained as long as we realise that it only applies to one special kind of non-contestation. In sum, on the basis of the evidence we have thus far examined the conclusion that Epicurus really held all alternative explanations to be true seems inevitable. Some reservations Those who maintain the truth of all alternative explanations (third option above), generally explain this truth by means of the principle of plenitude, referring to Lucretius’ testimony in DRN V 526-33 (item F on on p.19 above), which is usually taken to mean that, although for each individual event only one explanation can be true (though we do not know which one), with respect to the general type of event that is being explained every alternative explanation that is not contested by the appearances is true.69 I have a number of reservations about this claim, however. My first reservation concerns the meaning of the word ‘true’. Although the principle of plenitude provides a way in which each one of a number of alternative explanations may be called true, it must be observed that this truth is something very different from the universal and ubiquitous truth attaching to those theories which exhibit a singular agreement with the appearances, and – despite his unqualified claim that non-contestation establishes truth – Epicurus does seem to acknowledge the difference. Why else would he, in the introduction of his Letter to Pythocles (2 [86]; see item E on on p.18 above), oppose plural to singular agreement with appearances? In order to serve as the foundation of a systematic physical theory some tenets not only allow but actually require not just agreement but singular agreement with appearances, so as to be pronounced universally true, and this singular agreement with appearances can only be established by contestation of the contradictory hypothesis. It is almost as if Epicurus were saying that all explanations are true, but some (viz. singular explanations) are more true than others. In this respect Gisela Striker’s division of Epicurean non-contestation into a strong kind which establishes truth and a weak kind which establishes only possibility (second option above) is actually a good approximation of Epicurus’ use of non-contestation. My second reservation is about the general application of the principle of plenitude to all cases of multiple explanations. It seems to be universally agreed upon that Lucretius’ account of the principle may be generalised in this 69

Asmis (1984) 322, 324-25; id. (1999) 289; Sedley (1982) 270 with n.72; Long & Sedley (1987) vol.1, 95-96; Allen (2001) 197-198.



way, but in fact his account is appended to, and only explicitly refers to, the alternative explanations of the motions of the stars. Yet, even if we allow that his words have a broader application, it seems legitimate to investigate the scope of their applicability. In order to be able to apply the principle of plenitude each phenomenon under investigation needs to be viewed as an instance of a general type of events. In the case of meteorological phenomena this general type can be easily envisaged even without reference to the infinity of worlds. Thunder, for instance, is accounted for by Lucretius with nine different explanations.70 Only one of these will be true with respect to one particular thunderclap (although we do not know which one), but all of them are true with respect to thunder in general. In the case of astronomical phenomena, which are often concerned with unique objects, it is harder to accept them as instances of some general type. Only if we are prepared to accept ‘sun’ and ‘moon’ as generic terms for objects of which there may be only one in this world but infinetely many in the universe at large, can we claim that each possible explanation is also true. Yet, what are we to do with multiple explanations for exceptional local phenomena such as eruptions of the Etna, the summer flooding of the Nile, the anomalous daily temperature fluctuation of the spring near the shrine of Hammon, etcetera, which Lucretius discusses in the second part of book VI (see also p.99ff below)? Are we to suppose that in the universe at large there are infinitely many Niles flowing down from infinitely many Aethiopian mountain-ranges each overflowing in summer to irrigate infinitely many Egypts? That is, in fact, a logical outcome of the assumption of an infinity of worlds. Yet, if Lucretius had wished us to think of a general type of event, he would not have emphasized all these inessential particulars, but he would have spoken about rivers that overflow in summer, of which the Nile only presents the most notable example.71 It is clear, then, that here Lucretius was not thinking of general types of events, but of particular and in some cases even unique72 local phenomena, to which the ‘principle of plenitude’ does not apply. In these cases each of the alternative explanations can at best be called possible, not true.73 70

DRN VI 96-159. Cf. Epic. Pyth 18 [100]. Seneca, N.Q. III 26, 1, informs us that, according to Theophrastus, a number of rivers in Pontus showed this same behaviour. 72 In DRN VI 712-13 Lucretius calls the Nile unique (‘unicus’) for overflowing in summer (but see n.71 above). 73 It must be noted that for phenomena of this class Lucretius most often provides only one explanation, sometimes plus a subsidiary one. The summer flooding of the Nile, with four explanations, is exceptional in this respect too: see APPENDIX B on p.245 below. 71



There is another reason why in some of these cases the alternative explanations cannot all be called true. Whereas astronomical and meteorological phenomena cannot be physically approached, and therefore rightly belong to the class of non-evident things (êdhla) which are typically tested by contestation and non-contestation (see item A on p.13 above), some of the exceptional local phenomena described in DRN VI do not necessarily defy closer observation, and so may seem open to testing by attestation and non-attestation (see item A on p.13 and item B, first paragraph, on p.14 above). One could simply go to Egypt and observe whether the annual flooding of the Nile is somehow correlated to the onset of the etesian winds (715-23), or to the formation of sandbanks in the mouths of the river (724-28), or to the onset of seasonal rains upstream (729-34), or to the melting of snow in the Aethiopian mountains (735-37).74 So, rather than being true for not being contested by appearances, these phenomena may be said to be still waiting (prosm°nonta) to be attested or not by closer observation (see item A on p.13 above), and so be neither true nor false. In this respect the example Lucretius has chosen to illustrate the method of multiple explanations in book VI is very appropriate after all. In VI 703-11 (item G on p.20 above), immediately preceding the account of the Nile flood, Lucretius compares the use of multiple explanations to the procedure one should adopt when viewing a dead body from afar: since no cause of death can be excluded all causes should be accepted as possible. Yet, just like the Nile, a dead body can be examined at closer range, and so (to a certain extent) reveal the causes. In cases like these, then, we may not be justified in calling every alternative explanation true. In sum: all alternative explanations are true, (1) insofar as they concern non-evident phenomena which are subject to contestation and noncontestation, and (2) insofar as these phenomena are conceived of as instances of a general type of event, and even then we have to subscribe to a very meagre conception of truth, which common parlance would rather refer to as possibility. Gisela Striker’s interpretation (see option 2 above) turns out to be not so bad after all. I would like to add one final observation concerning the principle of plenitude. This principle, on which the truth of all alternative explanations rests, itself depends on the assumed existence of an infinite number of worlds, which in turn depends on the assumed existence of an infinite number of For the general scarcity of alternative explanations for phenomena of this class see the second paragraph of § 2.3.3 on p.113 below. 74 Or, to be more precise: one might have done so before the completion of the Aswan Dam in 1970 which effectively cancelled the Nile’s annual flooding.



atoms in infinite space. However, as we shall see in the final chapter (see p.189 below), Lucretius’ argument for an infinite number of atoms may not be as strong as he would have wished. Conclusion Although the conclusion that Epicurus considered all alternative explanations true follows logically from his use of non-contestation to support individual alternative explanations and his claim that non-contestation establishes truth, there turns out to be much that detracts from this conclusion. His efforts to provide a more certain basis than mere non-contestation for his fundamental physical theories show that he did not set much value on the truth of multiple explanations, and Lucretius’ failure to generalize the principle of plenitude to all instances of multiple explanations clearly shows the limitations of identifying possibility with truth. It need not surprise us therefore that Epicurus himself in the Letter to Pythocles most often speaks of the alternative explanations as being merely possible.75

1.3.3 Contestation The Epicureans’ criterion for rejecting theories is pretty straightforward: a theory is false if it disagrees with, or is contested by, the appearances (see item B on p.14 above). In practice Epicurus and Lucretius only rarely reject theories. Epicurus does, however, repeatedly mention one kind of explanations that should never be admitted in physical enquiry and especially astronomy and meteorology, viz. those explanations which attribute these phenomena to the involvement of the gods. Such an involvement, Epicurus holds, would be in conflict with the blessed nature of the gods and therefore must be rejected.76 Several examples of such theories are mentioned by Lucretius, who in VI 379422 argues against the popular view that thunderbolts are Jupiter’s work, in VI 753-4 against the myth that crows avoid the Athenian Acropolis because of Pallas Athena’s wrath, and in VI 762-6 against the belief that Avernian places are the gates to the Underworld. Yet, even explanations which do not rely on divine interference may sometimes be rejected. One example of this is, again, provided by Lucretius. In DRN VI 848-78 he discusses the curious behaviour of the spring near the shrine of Hammon, whose water is cold during the day and hot at night. Before embarking on his own account of the matter, he first describes and rejects a theory brought forward by ‘people’ (homines), who 75 76

See APPENDIX A on p.243 below. Epic. RS 1; Hdt. 76-7; 81; Pyth. 14 [97]; 33 [113]; 36 [115-6]. Cf. Lucr. II 1090-1104; V 156-234; 1183-1240; VI 50-79; 379-422; Cic. ND I, 52.



claim that the sun heats the spring from below during its nocturnal passage under the earth. Lucretius rejects this theory on the ground that, if the sun were able to affect the spring from below through the vast body of the earth, it would affect the spring even more when shining down on it unimpeded by the earth. But then the spring would have to be even hotter during the day, which is not observed to happen. Although the explanation is free from religious superstition, it fails to explain the actual phenomenon and therefore must be rejected. While for the Epicureans incompatibility with appearances is a sufficient and necessary ground for rejecting a theory, incompatibility with other theories is not. In an article written in 1978, Abraham Wasserstein faults the Epicureans for this.77 If the Epicureans had been truly committed to science, he writes, they should have paid attention to the fact that many of their theories are interdependent so that elimination of one theory may bring along the elimination of another. For instance, the theory that the sun is extinguished at night78 is incompatible with the theory that the moon receives its light from the sun,79 and consequently with those explanations of the moon’s phases and eclipses which presuppose that the moon shines with reflected light.80 Although Wasserstein’s observations are correct, I do not entirely agree with his point. It is true that in the Letter to Pythocles and in the astronomical and meteorological sections of the DRN phenomena are generally presented in isolation,81 but from an Epicurean-epistemological perspective it makes no difference for a theory’s possibility if it logically depends on another possible theory. It would only matter if this second theory turned out to be in conflict with appearances: in that case every theory that depends on it would share its downfall. What Wasserstein has shown, then, is not so much a flaw, as a vulnerability of Epicurus’ and Lucretius’ multiple explanations: if one explanation were to be eliminated – because of an inconsistency with appearances which E. and L. might have overlooked – this could lead to an avalanche of further eliminations, in some cases – where the explanations are exhaustive – even resulting in single explanations. 77

Wasserstein (1978) 490-4. Epic. Pyth. 7 [92] 1st explanation; Lucr. V 650-3, 660-2. 79 Epic. Pyth. 11 [94-5] 2nd explanation; Lucr. V 705. 80 Lucr. V 705-14 and 762-7. 81 A salient example is Epicurus’ and Lucretius’ separate treatment of thunder (Pyth. 18 [100]; DRN VI 96-159), lightning (Pyth. 19 [100-1]; DRN VI 160-218) and thunderbolts (Pyth. 21 [103-4]; DRN VI 219-422), as if they were three independent phenomena, rather than symptoms of a single phenomenon. On the other hand in the DRN explanations of lunar phases (705-50) and eclipses (762-70) are explicitly distinguished according to whether they assume that the moon shines with its own or with reflected light. 78



1.3.4 Non-contestation and analogy Until now I have managed to evade the question of how non-contestation establishes the possibility (or truth) of multiple explanations. Above I have argued for the existence of two kinds of truth; now it will be necessary to distinguish two kinds of possibility as well. If non-contestation is the absence of contestation, and if contestation of a theory consists in tracing a fatal incompatibility with the appearances (as Sextus explains), then noncontestation might be interpreted as the failure to trace such an incompatibility. In that case we would have established the theory’s subjective possibility: the theory is possible as far as our knowledge goes; it cannot be excluded that at some later point in time new information may force us to reconsider and reject the theory. However, Epicurus’ equation of possibility with truth clearly shows that he had in mind something more fundamental than that, viz. objective possibility, a possibility beyond the limitations of our knowledge, residing in the structure of the universe itself.82 However, in order to establish such an objective possibility we would need not just the absence of contestation, but the certainty that the theory will never be contested by appearances. How can such a certainty be obtained? In the astronomical and meteorological accounts of Epicurus and Lucretius an important role is assigned to analogy. The astronomical and meteorological sections of Lucretius’ DRN abound in specific analogies (as does the rest of his work).83 Almost every single explanation is illustrated by a specific analogy from everyday experience. Although, as a poet, Lucretius knows how to exploit these analogies poetically, the fact that many of his particular analogies are identical to those known from other, non-poetical, works on these subjects (like the Syriac meteorology84 (see p.64 below) and Seneca’s Naturales Quaestiones85) suggests that their primary role was scientific. Lucretius does not tell us what this role is, but we may perhaps learn more from Epicurus.


Allen (2001) 197. On Lucretius’ use of analogies see e.g. Schrijvers (1978) and Garani (2007). 84 Many of the parallels are noted in Daiber (1992) 272-82. The degree of correspondence is variously assessed: while Kidd (1992), 301, observes ‘close parallels including the illustrative analogies’ between Lucretius and the Syriac meteorology, Garani (2007), 97, instead notes ‘the remarkable lack of correspondence between Theophrastean [this is a reference to the Syriac meteorology] and Lucretian analogies.’ 85 See e.g. Bailey’s commentary on Lucretius book VI. 83



Although the number of specific analogies in the Letter to Pythocles is very limited86 (probably due to its being a summary),87 Epicurus does provide some useful theoretical remarks about the use of analogy in general. According to Epicurus (Hdt. 80), “we must carefully consider in how many ways a similar phenomenon is produced here with us, when we reason about the causes of the phenomena above as well as everything non-evident”, and (Pyth. 2 [87]) “signs of what happens in the sky can be obtained from some of the phenomena here with us: for we can observe how they come to pass, though we cannot observe the phenomena in the sky: for they may be produced in several ways”.88 Even more explicit is Pyth. 10 [94], on the phases of the moon, which, according to Epicurus, may be accounted for “in all the ways in which phenomena here with us, too, invite us to explanations of this appearance”.89 In these passages analogy seems to be presented as a heuristic device: its purpose is to provide signs, or to invite us to consider certain explanations.90 We might be tempted at this point to ascribe to Epicurus a scientific method consisting of two neatly distinguished stages, with analogy providing hypotheses, and non-contestation proving them.91 However, this is not the whole story. Sometimes analogy appears to be used not merely as a heuristic device, but as a proof in its own right.92 This can be seen e.g. in Pyth. 11 [95], where two alternative theories about the light of the moon are backed up in the following way: Ka‹ går par' ≤m›n yevre›tai pollå m¢n §j •aut«n ¶xonta, pollå d¢ éf' •t°rvn. ka‹ oÈy¢n §mpodostate› t«n §n to›w mete≈roiw fainom°nvn, §ãn tiw toË pleonaxoË trÒpou ée‹ mnÆmhn ¶x˙ ka‹ tåw ékoloÊyouw aÈto›w Ípoy°seiw ëma ka‹ afit¤aw sunyevrª ka‹ mØ énabl°pvn efiw tå énakÒlouya taËt' Ùgko› mata¤vw ka‹ katarr°p˙ êllote êllvw §p‹ tÚn monaxÚn trÒpon.


For here with us, too, we see many things having light from themselves, and many having it from something else. And nothing in the phenomena in the sky impedes this, if one always remembers the method of manifold causes and investigates hypotheses and explanations consistent with them, and does not look to inconsistent notions and emphasize them without cause and so fall back in different ways on different occasions on the method of the single cause. (tr. Bailey, modified, my emphasis)

In Pyth. 6 [91] the sun is compared to terrestrial fires (oÏtv går ka‹ tå par' ≤m›n purå ... yevre›tai), and in 18 [100] (1) thunder-production due to the wind whirling about in a hollow cloud is compared to a similar effect occurring in vessels (kayãper §n to›w ≤met°roiw égge¤oiw). 87 On the character and structure of Epicurus’ Letter to Pythocles see p.85ff below. 88 Translations by Bailey, with modifications. 89 Translation by Bailey, with modifications. A further example is found in Pyth. 19 [101]. 90 Allen (2001) 196-97. 91 Allen (2001) 197. 92 Allen (2001) 197.



The analogy with what happens ‘here with us’ is clearly presented as a proof. It is true that immediately afterwards Epicurus invokes noncontestation (‘nothing impedes’) as well, but significantly he restricts its use to ‘phenomena in the sky’ only, thereby suggesting that the ‘phenomena here with us’ have already been covered by the analogy. Apparently then, analogy with phenomena here with us implies agreement (sumfvn¤a) with phenomena here with us. If it is subsequently found that none of the phenomena in the sky contests either, i.e. if the explanations are not at variance with the original (celestial) object of inquiry, the explanations must be accepted as objectively possible. Other examples of this probative, as opposed to heuristic, use of analogy are found in Pyth. 6 [91], on the size of the sun (“for so too fires on earth ...”) and Pyth. 15 [98], on the length of nights and days (“as we observe occurs with some things on earth, with which we must be in harmony (sumfvn¤a) in speaking of celestial phenomena”). According to Epicurus, then, analogy performs two functions, a heuristic and a probative one.93 The first function logically precedes the second. If an explanation is needed, it must first be found, and then be verified. One and the same analogy may perform both functions, but not simultaneously. Once an explanation has been found, the analogy has performed its heuristic function, and can no longer serve in that capacity. Of course one may still report the specific analogy that led to the discovery of a certain theory, but such a report can no longer be called heuristic but at best historical and anecdotal. The actual heuristic use of analogy is therefore rather limited. It can never be linked with specific theories, which, after all, have been found already. Epicurus does seem to realise this. In those passages which appear to be dealing specifically with analogy in its heuristic capacity (Hdt. 80; Pyth. 2 [87] and 10 [94]), he never refers to specific explanations.94 We might of course still view Epicurus’ and Lucretius’ lists of alternative explanations as the outcome of an extensive heuristic use of analogy on Epicurus’ part, but even that isn’t exactly true. As will be demonstrated below (see §1.4 on p.53ff), almost all alternative explanations offered by Epicurus and Lucretius appear to derive from earlier thinkers. Not analogy, it turns out, but doxography seems to have been Epicurus’ favourite heuristic device.

93 94

See Allen (2001) 195ff At Pyth. 10 [94] ka‹ katå pãntaw kay' oÓw ... ktl. [“and in all ways in which” ... etc.] does not refer back to the three explicit explanations already given, but to other explanations that may at some point in the future be added to the list.



In explaining the many specific analogies in books V and VI of the DRN we may therefore disregard their heuristic function. The main purpose of these analogies is probative: they prove the possibility of an explanation.95

1.3.5 Degrees of probability and personal preferences Above we concluded that, at least with respect to a particular event, each one of a number of alternative explanations can at best be called possible. Yet, the question remains whether they are all equally possible. The only explicit statement on this subject is found in the Epicurean inscription of Diogenes of Oenoanda, fr.13 III 2-13, in the middle of a discussion of astronomical phenomena: TÚn zhtoËntã ti per‹ t«n édÆlvn, ín bl°p˙ toÁw toË dunãtou trÒpouw ple¤onaw, per‹ toËd° tinow mÒnou tolmhrÚn katapofa¤nesyai: mãntevw går mçllÒn §stin tÚ toioËton µ édrÚw sofoË. tÚ m°ntoi l°gein pãntaw m¢n §ndexom°nouw, piyan≈teron dÉ e‰nai tÒnde toËde Ùry«w ¶xei.

If one is investigating things that are non-evident, and if one sees that several explanations are possible, it is reckless to make a dogmatic pronouncement concerning any single one; such a procedure is characteristic of a seer rather than a wise man. It is correct, however, to say that, while all explanations are possible, this one is more plausible than that. (tr. Smith, slightly modified, my Italics)

The first part of this statement corresponds exactly to what we already know about Epicurus’ method.96 The last sentence, however, is not paralleled in any of Epicurus’ surviving works, nor in Lucretius’.97 Besides, there is something self-contradictory about Diogenes’ words, for, after denouncing as ‘seers’ those who opt for a single explanation, he himself seems to be singling out one explanation under the guise of plausibility. It would have been interesting to know on what grounds Diogenes would have us consider one explanation more plausible than another, but unfortunately he either failed to inform the reader, or the relevant part of the inscription is lost. As neither Epicurus nor Lucretius have left us any explicit theoretical considerations about the admissibility or inadmissibility of applying different degrees of probability, we cannot know for certain how Diogenes’ remark relates to Epicurean orthodoxy. In the absence of theoretical considerations any practical application of the principle would be of help too. In the Letter to Pythocles, however, Epicurus nowhere expresses a preference for any particular explanation. We do, however, have one 95

This does not mean that they may not perform other functions too; many of the analogies in Lucretius seem to have an illustrative function as well: they help the reader form a mental picture by providing a conceptual model. Besides they often provide Lucretius with an excellent excuse to show off his poetic genius. 96 For parallels in Epicurus and Lucretius, see Smith (1993), 455, n.8. 97 So Algra (2001) n.28.



testimony which might be interpreted as attributing to Epicurus just such a personal preference. In the Naturales Quaestiones, VI 20, 7, Seneca, having just reported a whole list of alternative explanations of earthquakes as brought forward by Epicurus, concludes with the following words: Nullam tamen illi {sc. Epicuro} placet causam motus esse maiorem quam spiritum.

No cause of an earthquake, however, Epicurus deems to be greater than wind.

One’s interpretation of these lines depends strongly on the meaning one wishes to attribute to maiorem, ‘greater’. One possible meaning in this context would indeed be ‘more likely’. Epicurus might have said that, although there are many possible causes, those involving wind are the most likely. However, as with most meteorological occurrences, an earthquake is not a single, recurrent, phenomenon, but every earthquake stands alone. It is perfectly possible, therefore, that one earthquake is brought about by cause A, while another is caused by B. Under these circumstances ‘more likely’ is almost equivalent with ‘more frequent’, which – in this context – is another possible interpretation of maiorem. So, Epicurus might have said that, although earthquakes can be, and are, produced by many different things, they are most often produced by wind. Yet, there is another possible interpretation. Different causes may have different effects. Maiorem in this context may also mean ‘more powerful’. Epicurus may have said that, although earthquakes can be, and are, produced by many different things, the strongest ones are produced by wind. That this is in fact the correct interpretation is borne out by the way in which Seneca continues (VI 21, 1): Nobis quoque placet hunc spiritum esse qui possit tanta conari, quo nihil est in rerum natura potentius, nihil acrius, sine quo ne illa quidem quae uehementissima sunt ualent.

We (i.e. the Stoics) too believe that it is this wind, which can attempt so much, which is mightier and fiercer than anything in nature, without which not even those things which are strongest have power.

So, according to Seneca, Epicurus held that wind is the most powerful cause of earthquakes. Like Seneca himself, Epicurus may have been brought to this view by Aristotle.98 Be that as it may, Seneca’s testimony cannot serve 98

Arist. Mete. II 8, 365b29 – 366a5: “Our next step should therefore be to consider what substance has the greatest motive power. This must necessarily be the substance whose natural motion is most violent. The subsance most violent in action must be that which has the greatest velocity, as its velocity makes its impact most forcible. The farthest mover must be the most penetrating, that is the finest. If, therefore, the natural constitution of wind is of this kind, it must be the substance whose motive power is the greatest. For even fire when conjoined with wind is blown to flame and moves quickly.



to confirm that Epicurus himself admitted different degrees of probability, or ever voiced personal preferences for any one of the alternative explanations. So much for Epicurus himself. As for Lucretius: there is one instance in the DRN where Lucretius seems to express a preference for a particular explanation. At V 621-2 he introduces his first explanation of the yearly and monthly motion of the sun and the moon with the following words: Nam fieri vel cum primis id posse videtur, Democriti quod sancta viri sententia ponit: …

For, in the first place it seems that this may be the case, what the sacred opinion of the man Democritus states: …

There is some ambiguity in the words ‘cum primis’. ‘Cum primis’ or ‘cumprimis’ literally means: with or among the first. This can in principle be interpreted in two ways: 1. Most often ‘cum primis’ or ‘cumprimis’ (like the synonymous ‘in primis’ or ‘imprimis’) is used to indicate that what is said is so in the highest degree, or particularly.99 On this interpretation ‘vel’ is best understood as an intensifying particle100 with ‘cum primis’: ‘among the very first’. In the present case this would mean that Democritus’ theory seems to be possible in the highest degree. This interpretation underlies the translations of e.g. Rouse & Smith (‘For among the most likely causes is that ...’) and Leonard (‘Yet chief in likelyhood seemeth the doctrine ...’). 2. Occasionally ‘cum primis’ or ‘cumprimis’ is used to indicate the first item in a series.101 (Its synonym ‘in primis’ or ‘imprimis’ is actually used in this sense quite often).102 Interpreted in this way the expression may be rendered as for a start or to begin with. ‘Vel’ may again be an intensifier with ‘cum primis’,103 or it may be used to imply “that other instances might be mentioned at will”.104 Used in this sense ‘vel’ may be rendered as ‘for instance’ or ‘for example’. This second interpretation is followed by e.g. Bailey (“For, first and foremost, it is clear that it may come to pass ...”) and Ernout (“Tout d’abord il est possible semble-t-il, que les choses se passent So the cause of earth tremors is neither water nor earth but wind, which causes them when the external exhalation flows inwards” (transl. Lee). 99 Lewis & Short ‘primus’ II B & ‘1. cum’ II D; OLD ‘cum1’ 6e & ‘primus’ 15c (cf. OLD ‘imprimis’ 1). 100 Lewis & Short ‘vel’ II B 1&2; OLD ‘vel’ 5c.Cf. Lucr. VI 1237 “vel in primis”. 101 Not in the OLD and Lewis & Short, but see Plaut. Truc. 660-1: “eradicarest certum cumprimis patrem,/ post id locorum matrem.” and Apul. Flor. 16, 36: “cum primis commemorauit inter nos iura amicitiae {…}; tunc postea uota omnia mea {…} recognouit.” TLG ‘cumprimis’ fails to distinguish different meanings. 102 OLD ‘imprimis’ 2. 103 See note 100. 104 Lewis & Short ‘vel’ II C. Cf. OLD ‘vel’ 4 a&b.



...”). See also Bailey’s comment at V 621: “This makes it clear that Lucr. intends to expound the first of a series of alternative causes {...}”. Which of these two interpretations is the right one? The first interpretation attributes to Lucretius a preference for Democritus’ view. Such a preference, however, seems to be unmotivated. The only way in which the present account differs essentially from other explanations in the astronomical and meteorological sections of DRN, is the explicit attribution to Democritus, whom Lucretius clearly admires. Yet admiring Democritus does not necessarily imply a preference for his theories: in III 370-3 another theory of Democritus, introduced with the same words of admiration (III 371 = V 622), is flatly rejected! It remains unclear, therefore, why in this case Democritus’ view should be considered to be among the most likely causes. Besides, such a preference for a single theory is also quite unprecedented in the DRN. Books V and VI contain scores of problems for which several alternative explanations are offered. Why should Lucretius in this one case express his preference, and nowhere else? Finally, such a preference seems also to be unwarranted by Lucretius’ own methodological remarks. Less than 100 lines earlier Lucretius stressed that out of several possible causes “one {...} is necessarily the case here, / {...} but which of them it is, / is not for them to lay down who proceed step by step.” (V 531-3).105 If Lucretius had thought it permissible to assign different degrees of probability to the alternative explanations this would have been the place to say so. But he did not. For these reasons I think this interpretation should be rejected. The second interpretation, although based on a less frequent use of the expression ‘cum primis’, provides a good alternative, which avoids all of the above problems. This brings us back to Diogenes of Oenoanda’s assertion that “it is correct {...} to say that, while all explanations are possible, this one is more plausible than that.” Not only is the claim itself without precedent in earlier Epicurean writings, but now our search for applications of this principle has yielded nothing either. It seems safe to assume, therefore, that Diogenes’ assertion is a later innovation. We do not know what the reason for this innovation was, nor how Diogenes himself applied it, but perhaps one example of its application can be unearthed from the ruins of his inscription. Diogenes’ claim is part of a fragment (fr.13) that begins with a promise to deal with the risings and settings of the sun, a problem for which Epicurus (Pyth. 7 [92]: see p.17 above) and Lucretius (V 650-79) had proposed two possible explanations. Unfortunately Diogenes’ fragment breaks off before he 105

For a fuller quotation, see above, p.19.



can deal with this specific problem. The same problem, however, is also discussed in another fragment (fr.66), where Diogenes criticises certain adversaries for “dismissing the unanimous opinion of all men, both laymen and philosophers, that the heavenly bodies pursue their courses round the earth both above and below ...” (tr. Smith). It is clear that Diogenes himself shares this ‘unanimous opinion of all men’, silently passing by Epicurus’ alternative explanation according to which the heavenly bodies are extinguished at night. It is possible that with other astronomical problems too Diogenes preferred to follow the generally accepted view, and he may have found this appeal to greater and lesser plausibility a convenient way to express these preferences without explicitly rejecting Epicurus’ alternative explanations, thus reconciling Epicurean orthodoxy with the accepted astronomical views of his time.106

1.3.6 Lucretius’ supposed preference for the theories of the mathematical astronomers. Although in the astronomical sections of Epicurus’ and Lucretius’ works we have found no explicit preferences for any one of a number of alternative explanations, there are some passages in the De rerum natura where Lucretius is believed to betray at least an implicit preference for a certain class of explanations. If true, this observation would contradict our earlier conclusion that Lucretius, like Epicurus, was impartial to the individual alternative explanations. It will be necessary therefore to deal with this claim as well. On p.58 of his commentary Bailey writes: “in the astronomical passages he [i.e. Lucretius] frequently places the right explanation first, as though he had a personal preference for it”. The point is repeated on p.1394, in the introduction to Lucretius’ astronomical section of V 509-770: “It should, however, be noticed that Lucr. usually places the true explanation first, as though he really preferred it.” Out of context, Bailey’s observation seems a bit trivial, for: who wouldn’t prefer the right and true explanation? In order to understand what Bailey really means by these terms, we must have a closer look at his comments on the individual sections of Lucretius’ astronomical passage. On p.1439, commenting on Lucretius’ discussion of the phases of the moon in V 705-50, Bailey writes: “This [i.e. the first] view {…} is clearly the view of the astronomers to whom Lucr. refers as his authorities in 713-14, and again as the astrologi in 728. {...} Lucr. therefore included it, and probably by placing it first meant to suggest that he believed it to be the right explanation 106

With respect to meteorological phenomena Diogenes seems to have had no qualms about offering alternative explanations: see fr.14 on the causes of hail, and fr.98.8-11 on the causes of earthquakes.



...”, and on pp.1446-7, commenting on Lucretius’ discussion of solar and lunar eclipses in V 751-70: “This [i.e. the first] theory, which was no doubt that of the ‘astronomers’ and is in fact the true explanation {…} is to be compared with the first theories put forward in 682-95 and 705-14. {…} Once again Lucr. by putting this theory first appears to give it the preference ...” Apparently then, when Bailey speaks of the ‘right’ and the ‘true’ explanation he means the explanation of the mathematical astronomers, which happens to be the true explanation. But Lucretius couldn’t have known that. That is precisely the point of his offering multiple explanations: that one cannot know for certain which is the right one.107 What he could have known, and did know, is that certain explanations came from the stock of the (mathematical) astronomers or ‘astrologi’ as he calls them. We may therefore rephrase Bailey’s observation as follows: “In the astronomical passages Lucretius frequently places the explanation of the mathematical astronomers first, as though he had a personal preference for it.” Now there are two sides to this observation: (a) the observed fact, and (b) Bailey’s interpretation of the fact. Let us first turn to the observed fact: “in the astronomical passages Lucretius frequently places the explanation of the mathematical astronomers first.” In the astronomical section Lucretius covers the following eleven subjects: 1. Motions of the stars (509-533) 2. Immobility of the earth (534-563) 3. Size of the sun, moon and stars (564-591) 4. Source of the sun’s light and heat (592-613) 5. Turnings of the sun, moon and planets (614-649) 6. Causes of nightfall (650-655) 7. Causes of dawn (656-679) 8. Varying lengths of day and night (680-704) 9. Phases of the moon (705-750) 10. Solar eclipses (751-761) 11. Lunar eclipses (762-770) Two of these subjects (2 and 3) do not, apparently, admit of more than one explanation, and the explanation given in each case is certainly not that of the mathematical astronomers. Lucretius explains the immobility of the earth on the assumption that it floats on a cushion of air, a view not easily reconciled with the spherical earth of mathematical astronomy (see Chapter 3 and esp. 107

See e.g. DRN V 526-33 (see text on p.19 above).



§3.3.6 below), and the heavenly bodies he claims to be the size they appear to be, which is usually interpreted as being very small, whereas the mathematical astronomers, for all their different estimates, at least agreed that the sun and the stars are much larger than the earth, and only the moon somewhat smaller.108 That leaves us with nine cases where we can test Bailey’s observation. The first of these, about the (daily?) motions of the stars, is problematic. Lucretius offers five possible explanations in all, which – however – fall into two main divisions: either (a) the whole sphere of the sky revolves, carrying the heavenly bodies along, or (b) the sky stands still, while the heavenly bodies move independently. The first option was – in fact – the view of the mathematical astronomers.109 Lucretius, however, goes on to subdivide these main divisions, offering two possible physical explanations for the first option, and three for the second, in a way that goes beyond the constraints of mathematical astronomy, which only concerned itself with the mathematical, i.e. quantitative and geometrical, aspects of astronomy.110 It remains unclear, therefore, whether we should consider this passage a case in point. Bailey himself does not seem to have viewed it as such, for in his commentary to this passage he makes no reference to his own observation. The next subject where a plurality of explanations comes into play is item 4, on the source of the sun’s light and heat. Lucretius offers three explanations, none of which can be related to the mathematical astronomers. In fact the ancient astronomers have left us no view on this subject at all, which falls outside the scope of their competence, i.e. the quantitative and geometrical aspects of astronomy (see above). The following subject (5), the turnings of the sun, moon and planets, falls well within the competence of the mathematical astronomers and we know what their solution to this problem was. According to the astronomers, the sun, the moon and the planets exhibit a slower, secondary, east-ward motion on top of the daily, west-ward, revolution of the fixed stars. This secondary motion they all perform more or less along the same circular path, the socalled zodiac. The zodiac is not parallel to the equator but inclined to it by an angle of about 23.5°. This fact was referred to in antiquity as the obliquity of the zodiac – ≤ lÒjvsiw toË zƒdiakoË / obliquitas signiferi (sc. orbis / 108

Cf. Cleomedes II 1-3 (esp. II 3.68ff). See also Heath (1913) 328-350, esp. the tables on p.332 and p.350. 109 Evans (1998), p.75 with nn.3 and 4. 110 See e.g. Arist. Ph. II 2, 193b22ff; Stoics apud D.L. VII 132; Posidonius F18 E-K (= Geminus apud Alexandrum apud Simplicium In phys. 291.21-292.21); Sen. Ep. 88.2528; Strabo 1.1.20 & 2.5.2; etc. For modern views on the matter see e.g. Bowen-Todd (2004), 6 & 193-9, and Evans (1998), 217-219.



circulus). In Epicurus’ own treatment of the subject, in the Letter to Pythocles 9 [93], this explanation is the first to be mentioned.111 Lucretius too shows himself to be aware of the existence of this explanation, for, in a later passage (8), when discussing the related problem of the varying lengths of day and night, he clearly alludes to it.112 In the present passage, however, he doesn’t say a word about it. So, far from placing the mathematical astronomers’ view first, Lucretius chooses to ignore it. Instead, as we have seen, he starts with an explanation explicitly ascribed to Democritus. The first instance where Lucretius does include the view of the mathematical astronomers among a number of possible explanations is subject 6, on the causes of nightfall. Night ensues, he says, either because the sun, upon reaching the westernmost point of its orbit, is extinguished, or because the sun, upon reaching this point, passes out of sight below the plane of the earth. Once again Lucretius does not conform to Bailey’s observation: the view of the mathematical astronomers is presented second, not first. The next subject (7), on the causes of dawn, is the mirror image of the previous subject. Again Lucretius gives us two explanations, which correspond chiastically to the two possible causes of nightfall: either the same sun, having reached the easternmost limit of its orbit, emerges again above the plane of the earth, or a new sun is born from small fires which collect in the eastern sky each morning. This time, at last, the view of the mathematical astronomers is presented first. The rest of the subjects (8-11) also follow this pattern, thereby conforming to Bailey’s observation. Mathematical astronomers explained (8) the seasonal variation of the length of nights and and days with reference to the sun’s position in the slanting zodiacal belt, (9) the phases of the moon with reference to the relative positions of the sun and the moon – assuming that the latter shines with light reflected from the former –, (10) solar eclipses as the moon blocking the sun from our view, and (11) lunar eclipses as the moon falling into the earth’s conical shadow and so being deprived of the sun’s light. In Lucretius’ account each of these theories is the first of a number of alternative explanations. The following table sums up our findings:


Tropåw ≤l¤ou ka‹ selÆnhw §nd°xetai m¢n g¤nesyai katå lÒjvsin oÈranoË oÏtv to›w xrÒnoiw kathnagkasm°nou ... 112 DRN V 691-3: ... propter signiferi posituram totius orbis, / annua sol in quo concludit tempora serpens, / obliquo terras et caelum lumine lustrans ...



Table 1-1: Place of the mathematical astronomers’ explanations in the astronomical section of DRN Passage


509-533 534-563 564-591 592-613 614-649 650-655 656-679 680-704 705-750 751-761 762-770

1. Motions of the stars 2. Immobility of the earth 3. Size of the sun, moon and stars 4. Source of the sun’s light and heat 5. Turnings of the sun, moon and planets 6. Causes of nightfall 7. Causes of dawn 8. Varying lengths of day and night 9. Phases of the moon 10. Solar eclipses 11. Lunar eclipses

Number of explanations 5 1 1 3 2 2 2 3 4 3 3

View of the mathematical astronomers 2 1 1 1 1 1

Only 5 out of 11 cases seem to conform to Bailey’s thesis. If, however, we confine ourselves to those subjects where the view of the mathematical astronomers is included at all, the ratio becomes 5 to 6, which seems significant enough. We may safely conclude therefore that the explanations of the mathematical astronomers were somehow privileged, although we do not yet know why or in what way. In only two of these cases the mathematical astronomers are explicitly mentioned or implied in a meaningful way. In lines 694-5, at the end of his first explanation of (8) the variation of day-length, Lucretius speaks of those “who have mapped the places of the sky, / all adorned with stars properly arranged”,113 which is a clear reference to the mathematical astronomers, who typically demonstrated their theories by means of celestial globes and planetaria.114 In lines 727-8, at the end of the third explanation of (9) the phases of the moon Lucretius speaks of the “art of the astronomers” (astrologorum artem), by which he appears to be referring back to the first explanation of that section. Apparently Lucretius assumes his reader to be familiar enough with contemporary astronomy to recognize this reference. In the same way the reader may be assumed to recognize the other, unidentified, references to the mathematical astronomers as well. 113 114

DRN V 694-5: ... qui loca caeli / omnia dispositis signis ornata locarunt. Plato, Ti. 40c-d, claims that the planetary motions can be properly demonstrated by means of a visible model only, and Epicurus, Pyth. 9 [93] (cf. On nature XI fr.38 Arr. with Sedley (1976), 32, 37-39), seems to associate the use of such models with mathematical astronomy. On the use of visibile model in astronomy see Cornford (1937) 74-6; and Evans (1998) 78-84.



According to Bailey, the privileged position of the mathematical astronomers’ explanations indicates that Lucretius himself preferred these over the other views, believing them to be the right ones. To Bailey this is so self-evident that he doesn’t even bother to defend this assertion. In fact, his claim is quite unfounded. Not only is such a preference, as we have seen, hard to reconcile with Lucretius’ insistence that all explanations offered have an equal claim to the truth (526-33), but it actually fails to take into account certain clues provided by Lucretius himself in several of the relevant passages. The first of these clues is at 713-14, where Lucretius concludes his first explanation of the phases of the moon (i.e. the explanation of the mathematical astronomers) with the following words: ut faciunt, lunam qui fingunt esse pilai consimilem cursusque viam sub sole tenere.

as they hold, who imagine the moon to be like a ball and to keep the path of her course below the sun.

Throughout the DRN the verb ‘fingere’ is used to stress the unfoundedness and even falsehood of theories,115 which are subsequently rejected. It would be most surprising if Lucretius would now use this same verb to refer to a theory which, in Bailey’s words, “he believed [..] to be the right explanation.” Another clue, which will require a bit more work, but may also help us to find the real reason why Lucretius gives priority to the explanations of the mathematical astronomers, is found at the end of the astronomical passage (lines 751-770), where Lucretius discusses the causes of (10) solar and (11) lunar eclipses. In his discussions of other astronomical phenomena Lucretius simply enumerates his multiple explanations, saying something like: ‘phenomenon X may be caused by A, or B, or C, etc.’, where A most often represents the view endorsed by mathematical astronomy. However, in his discussions of solar and lunar eclipses Lucretius employs a slightly different structure. This is Lucretius’ account of solar eclipses (753-761):


See e.g. I 371, I 842, I 847, I 1083, II 175 and V 908.




nam cur luna queat terram secludere solis lumine et a terris altum caput obstruere ei, obiciens caecum radiis ardentibus orbem,

For why should the moon be able to shut off the earth from the sun’s light and obstruct the sun’s high source from the earth, by interposing her dark orb to his burning rays,


tempore eodem aliut facere id non posse putetur corpus, quod cassum labatur lumine semper?

and not at the same time some other body, which always glides with unseen light, be thought able to achieve this?


solque suos etiam dimittere languidus ignis tempore cur certo nequeat recreareque lumen, cum loca praeteriit flammis infesta per auras, quae faciunt ignis interstingui atque perire?

and why could not the sun at a certain time from wearyness dismiss his fires and then again renew his light, when he has passed the regions harmful to his flames, which make his fires go out and die?

And this is how Lucretius explains lunar eclipses (762-770): A

et cur terra queat lunam spoliare vicissim lumine et oppressum solem super ipsa tenere, menstrua dum rigidas coni perlabitur umbras,

And why should the earth in turn be able to rob the moon of light, and keep the sun oppressed, being herself above, while in its monthly course the moon glides through the rigid shadows of the cone,


tempore eodem aliud nequeat succurrere lunae corpus vel supra solis perlabier orbem, quod radios inter rumpat lumenque profusum?

and not at the same time some other body be able to pass beneath the moon or glide above the solar orb, to interrupt his rays and flood of light?


et tamen ipsa suo si fulget luna nitore, cur nequeat certa mundi languescere parte, dum loca luminibus propriis inimica per exit?

and if, after all, the moon shines of herself with her own light, why could she not grow faint in a certain part of heaven, while passing through regions hostile to her own light?

For each of the two phenomena Lucretius offers three possible explanations, but, instead of simply enumerating them, as he usually does, he now marshals them into the format of a rhetorical question. The structure is the same in both cases: ‘Why should the (solar/lunar) eclipse be caused by ‘A’, and not by ‘B’ or ‘C’?’ The implied answer is, of course, that there is no good reason to prefer A over B and C. (So much for Bailey’s interpretation!) Yet, the way the question is put also suggests something else. Lucretius seems to be particularly worried that someone might consider ‘A’, i.e. the mathematical astronomers’ view, the only possible explanation. That someone might single out ‘B’ or ‘C’ in this manner does not seem to worry him at all! Why is that? Is it intrinsically worse to accept the view of the astronomers rather than any of the other theories? That is not what Lucretius is saying. Throughout the astronomical passage he has insisted upon the equal plausibility of each alternative explanation: no explanation is better or worse than any other. So, why then does Lucretius single out the explanation of the mathematical astronomers? An account with a somewhat similar rhetorical structure is found in the astronomical passage of Epicurus’ Letter to Pythocles. Here, in chapter 32 [112] Epicurus writes:


CHAPTER ONE Tinå str°fetai aÈtoË, ˘ sumba¤nei

Some stars revolve in their place, which comes to pass


oÈ mÒnon t“ tÚ m°row toËto toË kÒsmou •stãnai, per‹ ˘ tÚ loipÚn str°fetai, kayãper tin°w fasin,

not only because this part of the world is stationary and round it the rest revolves, as some say,


éllå ka‹ t“ d¤nhn é°row ¶gkuklon aÈto›w periestãnai, ∂ kvlutikØ g¤netai toË peripole›n …w ka‹ tå êlla,

but also because a whirl of air is formed in a ring round it, which prevents their moving about as do the other stars,


µ ka‹ diå tÚ •j∞w m¢n aÈto›w Ïlhn §pithde¤an mØ e‰nai, §n d¢ toËtƒ t“ tÒpƒ §n ⁄ ke¤mena yevre›tai,

or else it is because there is not a succession of appropriate fuel for them, but only in this place in which they are seen fixed,

ka‹ kat' êllouw d¢ ple¤onaw trÒpouw toËto dunatÚn suntele›syai, §ãn tiw dÊnhtai tÚ sÊmfvnon to›w fainom°noiw sullog¤zesyai.

and there are many other ways in which this may be brought about, if one is able to infer what is in agreement with appearances. (tr. Bailey, slightly modified)

This passage is concerned with the problem (not discussed by Lucretius) why some stars (e.g. those of Ursa Major and Minor) never set but revolve in their place. The first option, ‘A’, corresponds to what the mathematical astronomers said. Although Epicurus – like Lucretius – is normally perfectly happy to present his alternative explanations in the form of an uncomplicated disjunction (‘either A or B or C’),116 this time he has chosen a slightly more complex formulation: ‘not only A, but also B and C’. This formulation implies that Epicurus in reality only contemplates the possibility that someone might say ‘A’, not that someone might say ‘B’ or ‘C’. This is confirmed by the words “as some say” («kayãper tin°w fasin»), which Epicurus adds to explanation A. Apparently, explanation A had some actual support, which B and C, as far as we are told, had not. Something similar seems to be the case with Lucretius’ account of eclipses. In order to confirm this I will try to establish the extent of the contemporary support for view A, and the lack of such support for the alternative views, B and C. I will start with Lucretius’ account of lunar eclipses (V 762-770). A useful piece of information is provided by Aëtius, who in his chapter on the phases and eclipses of the moon (II 29), also reports a view that can be identified with Lucretius’ first explanation of lunar eclipses (together with an account of the phases):


See note 53 above, and the text thereto. For the various connectors used by Epicurus to articulate his lists of alternative explanations, see APPENDIX A on p.243ff below.

MULTIPLE EXPLANATIONS Yal∞w ÉAnajagÒraw Plãtvn ÉAristot°lhw ofl Stviko‹ to›w mayhmatiko›w sumf≈nvw tåw m¢n mhnia¤ouw épokrÊceiw sunodeÊousan aÈtØn ≤l¤ƒ ka‹ perilampom°nhn poie›syai, tåw d' §kle¤ceiw efiw tÚ sk¤asma t∞w g∞w §mp¤ptousan, metajÁ m¢n émfot°rvn t«n ést°rvn ginom°nhw, mçllon d¢ t∞w selÆnhw éntifrattom°nhw.


Thales, Anaxagoras, Plato, Aristotle and the Stoics (declare) in agreement with the mathematical astronomers that it (the moon) produces the monthly concealments by travelling together with the sun and being illuminated by it, and the eclipses by descending into the shadows of the earth which interposes itself between the two heavenly bodies, or rather when the moon is obstructed (by the earth).117

According to this report, Lucretius’ first explanation, which Bailey attributed to the astronomers, was also accepted by Thales, Anaxagoras, Plato, Aristotle and the Stoics. Thales and Anaxagoras, who had been long dead and left no schools to continue their thought, are irrelevant for the present purpose, but Plato and Aristotle, whose teachings were still followed in Lucretius’ time, and the Stoics, who had become the most influential philosophical sect of the period, are very relevant. From Aëtius’ report, which is confirmed by many other sources,118 it appears that Lucretius’ first explanation was the view, not just of the mathematical astronomers, but of every major school of philosophy still in existence in Lucretius’ day, Epicureanism excepted. The second and third explanations, on the other hand, do not seem to have been entertained by anyone later than Anaxagoras, who believed that lunar eclipses were also caused by interposition of other, unseen, heavenly bodies beside the earth, and Xenophanes, who ascribed all such phenomena to extinction and rekindling.119 Much the same can be said about Lucretius’ treatment of solar eclipses (V 753-561): his first explanation can again be attributed, not just to the mathematical astronomers, but to Aristotle and the Stoics and probably Plato


I have basically followed the reconstruction offered by Mansfeld & Runia (2009a), 61323, although I have opted for Stobaeus’ lectio difficilior «to›w mayhmatiko›w sumf≈nvw» (‘in agreement with the mathematical astronomers’) instead of Plutarch’s «ofl mayhmatiko‹ sumf≈nvw» (‘and the mathematicians in agreement’), a variation which M&R do not comment upon. The translation is freely adapted from M&R to suit my different reading of the text as well as my personal taste. 118 The attribution of this theory to Plato cannot be verified from his own works, but that he accepted the view that the moon is illuminated by the sun is clear from Resp. X 617a (with Heath (1913) 158); cf. Cra. 409a-b. Aristotle refers to the theory in Cael. II 14, 297b24-31, Mete. II 8, 367b20-22; Metaph. VIII 4, 1044b9-15; An. Post. I, 31, 87b39a2; II 2, 90a15-18; et passim. For the Stoics see e.g. SVF I 119, 120; II 676, 678 and Cleomedes II 6. 119 For attempts to identify Lucretius’ theories with those of his predecessors see the various commentaries.



too,120 while the second cannot be related to anyone at all, and the third to noone later than Xenophanes.121 In sum, Epicurus in the passage just quoted and Lucretius in his account of solar and lunar eclipses both start with the view of the astronomers, because that was the view that most people, including the other major philosophical schools, believed to be uniquely true. To this view Epicurus and Lucretius oppose other views that may not be current and popular, but which they consider equally possible. It seems reasonable to suppose that the same pattern applies also to the other cases in the astronomical section of the DRN where Lucretius starts with the view of the astronomers: (7) The view that dawn is caused by the same sun re-emerging from below the horizon can safely be attributed to Plato, Aristotle and the Stoics,122 all of whom conceived of the heavenly bodies as permanent entities, and we even have the explicit statement of a later Epicurean, Diogenes of Oenoanda (fr.66), that this was ‘the unanimous opinion of all men, both laymen and philosophers’. On the other hand, the alternative view that the sun is rekindled every morning can at best be assigned to Xenophanes, and perhaps Heraclitus and Metrodorus of Chios as well, but to no one later.123 (8) The theory that the annual variations in day-length are caused by the obliquity of the zodiac was at least maintained by the later Stoic Cleomedes,124 while the theory of the obliquity of the zodiac as such is attributable to Plato,


Plato, Ti. 40c-d, rightly attributes (solar and lunar?) eclipses to the interposition of another heavenly body (see also Cornford (1937) 135-6); for Aristotle see Div. Somn. 1 462b28-29; for the Stoics see SVF I 119; II 650 etc. 121 For attempts to identify Lucretius’ theories with those of his predecessors see the various commentaries. 122 For the Stoics see also Ar. Did. fr.32 (= SVF II 683) and Cleom. II 1.426-466. 123 Xenophanes A32, A33, A38, A40; Heraclitus B6; Metrodorus of Chios A4 D-K. On the other hand, the theory that the sun is quenched and rekindled is explicitly rejected by the Peripatetic Eudemus apud Theon phil. Expos. 199.21-22; and the Stoic Cleomedes II 1.426-466; further by Ptol. Alm. I 3, 11.24 – 12.18 (= I 1, 12 Heiberg); and Theon math. Comm. in Ptol. synt. 340-1 (Rome). 124 Cleom. I 3.76 – 4.1-17. See also Gem. 6,29ff, Vitruv. IX 3, 1-3 and Plin. N.H. II 17, 81.



Aristotle and the Stoics in general.125 The two alternative views, on the other hand, do not seem to have been held by anyone at all.126 (9) The section on the phases of the moon presents a slightly different story. Lucretius offers four alternative explanations. The first explanation – according to which the moon is illuminated by the sun, and the phases result from the changing relative positions of the two bodies – is easily recognized as the view of the mathematical astronomers.127 Lucretius does not, at this point, explicitly identify them, apart from the vague reference to those “who imagine the moon to be like a ball”.128 Yet, the use of present tense and plural already suggests that the theory did at least have some advocates in Lucretius’ day. In this particular case, however, the astronomers’ view seems to have met with a more serious challenge: at the end of his third explanation – according to which the moon is a sphere, one half of which is fiery, and which by revolving around its own axis produces the phases – Lucretius writes (727730)129: ut Babylonica Chaldaeum doctrina refutans astrologorum artem contra convincere tendit, proinde quasi id fieri nequeat quod pugnat uterque aut minus hoc illo sit cur amplectier ausis.

as the Babylonian doctrine of the Chaldeans, refuting the science of the astronomers, strives to uphold against them; just as if that which each of them fights for could not be, or as if there were less reason to embrace this than that.

The Babylonian theory is presented differently from the other two alternative explanations. Whereas the second and fourth explanations are, as we have come to expect, mere museum pieces, the view of the Chaldeans still 125

Ascription to Plato and Aristotle in Aët. II 23.5, and to the Stoics in SVF I 542, II 650.3 & 651.5 (= D.L. VII 144.3 & 155.9). For Plato see also Ti. 36b-d & 38e-39a (with Heath (1913) 159-60; or Cornford (1937) 72ff). For Aristotle see Gen. Corr. I 10.336a32-b24, 337a8 and Metaph. XII 5.1071a16 & 8.1073b17ff. The theory is also described by Aratus 525-544 and Vitr. IX 1,3. 126 See the various commentaries ad loc. 127 Aëtius II 29 (see p.47 above) ascribes the theory to the mathematical astronomers, Plato, Aristotle and the Stoics. The attribution of this theory to Plato cannot be verified from his own works, but he certainly accepted the view that the moon is illuminated by the sun (see n.118 above). Aristotle alludes to the astronomical theory of the lunar phases in Cael. II 11, 291b18-21 and An. post. I 13, 78b4-11. For the Stoics cp. D.L. VII 145 (= SVF II 650). Later Stoics (perhaps from Posidonius onwards: see Bowen-Todd (2004) 138 n.8, 141 n.19) held a slightly different theory, maintaining that the moon, on the side where it is touched by the rays of the sun, responds by emitting its own light: see Cleom. II 4, 21-32. On this later Stoic view see also p.49 (following n.130) below. 128 DRN V 713-4: lunam qui fingunt esse pilai / consimilem ... 129 The language used to describe this clash of opinions (esp. ‘convincere’) is the language of a law court. For other instances of conflicting views being described in judicial terms cf. Epic. Pyth. 10 [94] épodokimãz˙ (Bailey (1926): ‘put out of court’), Sen. NQ IVB 5.1 ‘litigant’ (Corcoran: ‘pleading in court’), and Hor. AP 78 ‘grammatici certant, et adhuc sub iudice lis est’.



seems to have been able to muster some real support among Lucretius’ contemporaries and later. This is confirmed by several other sources. In Vitruvius’ De architectura IX 2, written several decades after the death of Lucretius, two different theories of the phases of the moon are presented. The first, which corresponds to Lucretius’ third explanation, is attributed to Berosus the Chaldean, and the second, which corresponds to Lucretius’ first explanation, is attributed to the mathematician Aristarchus of Samos. It is significant that Vitruvius does not choose between the two. The same impartial attitude towards these two explanations we also encounter in Apuleius’ De deo Socratis 1.14-30 and Augustine’s Enarratio in Psalmos 10, 3. According to Augustine, both theories are probable, but it is humanly impossible to know which one is true. A different approach is found in the work of the Stoic Cleomedes, who may have lived some time around 200 AD.130 In II 4.1, Cleomedes discusses no less than three different theories concerning the phases of the moon. The first is again that of Berosus, the second the traditional view of the astronomers and the Peripatetics,131 and the third a Stoic modification of the former, according to which the moon, on the side where it is touched by the rays of the sun, responds by emitting its own light. As a Stoic, Cleomedes of course opts for the third alternative, but what is significant here is that he feels compelled to refute not only the theory of traditional astronomy, but also that of Berosus, as if both theories were equally relevant. We may assume therefore that Berosus’ theory was widely regarded as a reasonable alternative to the view of the astronomers, and one which could not be discarded as easily as other alternative theories. Lucretius shows himself to be aware of this. He chooses, however, to stick to his usual pattern, starting with the theory of the astronomers, and leaving the competing view of the Chaldeans for later. We have now established that Lucretius, when choosing to include the explanation of the mathematical astronomers, usually mentions it first. By doing so, Lucretius demonstrates both his awareness of the predominant position of their theories, and his determination to combat this predominance by pointing out that other solutions, whether newly invented or long-forgotten or still in vogue (like the Chaldean theory of lunar phases), are just as plausible. In the discussion above we have also examined one passage in Epicurus’ Letter to Pythocles, viz. ch.32 [112], which – like Lucretius, and for probably 130 131

For Cleomedes’ dates, see Bowen-Todd (2004) 2-4. See n.127 above.



similar reasons – starts with the view of the astronomers. It would be interesting to see whether in Epicurus’ Letter this is a sustained practice, like in Lucretius’ astronomical passages, or just an isolated case. In the following table I have set out the astronomical subjects of Epicurus’ Letter to Pythocles (excluding chapters 31 [111] on comets and 35 [114-5] on shooting stars, which, in antiquity, were not generally considered astronomical), each with the number of explanations given,132 and the place of the astronomers’ explanation, if included: Table 1-2: Place of the astronomers’ explanations in the astronomical passages of the Letter to Pythocles Chapter


7 [92] 8 [92-3] 9 [93] 10 [94] 11 [94-5] 12 [95-6] 13 [96-7] 14 [97] 15 [98] 32 [112] 33 [112-3] 34 [114]

Risings and settings Motions of the stars Turnings of the sun and moon Phases of the moon Origin of the moon’s light Face in the moon Solar and lunar eclipses Regularity of periods Varying lengths of day and night133 Stars turning in their place Planets Stars lagging behind

Number of explanations 2 3 4 3+ 2 2+ 2 1 2 3+ 2 3

View of the mathematical astronomers 2 1 1 2 2 2 1 1 2

The total number of astronomical subjects in the Letter is 13. In 9 cases we find the view of the astronomers included among a number of possible explanations.134 In only 4 of these cases the view of the astronomers is presented first. The general pattern Bailey had detected in the astronomical passages of the DRN does not seem to apply to the corresponding portion of Epicurus’ Letter to Pythocles. Yet, Lucretius’ critical attitude towards the views of the astronomers is not unlike Epicurus’. The Letter to Pythocles contains two explicit references to the (mathematical) astronomers (both in chapters where their view is presented first): in chapter 9 [93] Epicurus says: 132

A ‘+’-sign after the number indicates that Epicurus explicitly tells us that there may be still more explanations. 133 Sections 16-31 are devoted to subjects which traditionally belonged to meteorology and are therefore irrelevant for the present subject. 134 For identification of the individual explanations see the commentaries to the Letter to Pythocles.



Pãnta går tå toiaËta ka‹ tå toÊtoiw suggen∞ oÈyen‹ t«n §narghmãtvn diafvne›, §ãn tiw ée‹ §p‹ t«n toioÊtvn mer«n §xÒmenow toË dunatoË efiw tÚ sÊmfvnon to›w fainom°noiw ßkaston toÊtvn dÊnhtai §pãgein, mØ foboÊmenow tåw éndrapod≈deiw éstrolÒgvn texnite¤aw.

For all these and kindred explanations are not at variance with any clear-seen facts, if one always clings in such departments of inquiry to the possible and can refer each point to what is in agreement with appearances without fearing the slavish artifices of the astronomers. (tr. Bailey, slightly modified, my Italics)

And in chapter 33 [113]: TÚ d¢ m¤an afit¤an toÊtvn épodidÒnai, pleonax«w t«n fainom°nvn §kkaloum°nvn, manikÚn ka‹ oÈ kayhkÒntvw prattÒmenon ÍpÚ t«n tØn mata¤an éstrolog¤an §zhlvkÒtvn ka‹ efiw tÚ kenÚn afit¤aw tin«n épodidÒntvn, ˜tan tØn ye¤an fÊsin mhyamª leitourgi«n épolÊvsi.

But to assign a single cause for these phenomena, when the appearances demand several explanations, is madness, and is quite wrongly practised by persons who are partisans of the foolish notions of astronomy, and who give futile explanations of the causes of certain phenomena, whenever they do not by any means free the divine nature from the burden of responsibilities. (tr. Bailey, slightly modified, my Italics)

In his article ‘Epicurus and the mathematicians of Cyzicus’,135 David Sedley argues that such references should be viewed in the light of Epicurus’ rather personal feud with the Eudoxan school of mathematics and astronomy at Cyzicus. Although I agree that such personal animosity may certainly have added to the vehemence of Epicurus’ attacks, I think that here these attacks must be seen as having a broader application. In Pyth. 33 [113] Epicurus speaks of t«n tØn mata¤an éstrolog¤an §zhlvkÒtvn: ‘the partisans of the foolish notions of astronomy’. This is certainly an odd way to refer to just the astronomers, let alone such a specific group of astronomers. In fact the expression applies just as well, if not more, to all those who, while not being astronomers themselves, passionately embraced their findings, like e.g. Plato and Aristotle and their followers, and later the Stoics. At any rate, the second part of Epicurus’ criticism, that ‘they do not by any means free the divine nature from the burden of responsibilities’, applies more naturally to these philosophers, who made the gods responsible for the heavenly motions, than to the mathematical astronomers.136 135 136

Sedley (1976) 26-43; see esp. p.43 above. In Cic. ND I 30-39 the Epicurean spokesman Velleius explicitly criticises, among others, Plato (30), Aristotle (33), Theophrastus (35) and the early Stoics (36-9) for assigning divinity to the heavens and the heavenly bodies. For Plato’s views see e.g. Ti. 40a-d; Resp. VI 508a; Leg. VII 821b-c, X 899a-b, XII 950d; Epin. 981e, 983a-b, 984d (cf. Barnes (1989) 41); for Aristotle’s see e.g. Metaph. XII 8, 1074a38-b14, Cael. I



In Pyth. 33 [113] the ‘partisans of the foolish notions of astronomy’ are attacked, among other things, for assuming a single explanation, when the appearances call for several. The same criticism occurs throughout the Letter to Pythocles,137 but only here the target is specified. Yet, it is very likely that in the other instances too Epicurus was thinking in particular about these devotees of astronomy. Perhaps Anaxagoras in his time had been a proponent of single causes, or Empedocles, or Democritus, but that was a long time ago. In Epicurus’ days the only advocates of single explanations to be reckoned with were the astronomers and their partisans. In this respect Epicurus seems to have had the same reasons as later Lucretius for attacking their views, and Lucretius turns out the be firmly rooted in Epicurus’ track. Bailey’s observation that Lucretius in his astronomical passage usually presents the view of the astronomers first, appears to be basically correct. However, Bailey’s interpretation of the fact is wrong. Far from actually preferring the views of the astronomers, as Bailey supposed, Lucretius, like Epicurus before him, singles them out as the principal representatives of the wrong attitude towards the explanation of the non-evident. While the appearances call for several explanations and do not permit us to choose between them, the astronomers and their followers idly opt for a single explanation.

1.4 Multiple explanations and doxography Although, as Epicurus claimed, the appearances themselves invite us to adopt certain explanations,138 many of Epicurus’ and Lucretius’ alternative explanations were actually derived from earlier thinkers. For instance, to take our stock example, the theory that the sun is extinguished at sunset and rekindled at sunrise, which Epicurus and Lucretius consider a viable alternative for the common view that it passes unaltered below the earth, can be confidently identified with the view espoused by Xenophanes.139 Although the sources of individual alternative explanations are generally left unspecified, Lucretius does occasionally identify them: Democritus (V 621-2), the Chaldeans (V 727) and the ‘astrologi’ (V 728).140 We have also seen that in his astronomical section Lucretius often consciously (although mostly without explicit reference) starts his lists of alternative explanations with the 3.270b6-25, 9.278b14-16, II 1.284a12-14 & 284b3-5, 3.286a10-13, 12.292b32-293a1; for the Stoics see e.g. Cic. ND I 36-39, II 39, 42, 44, 54, 80. 137 Pyth. 2 [87], 10 [94], 15 [98], 33 [113], 34 [114]. See also Hdt. 80.5-6. 138 Epic. Pyth. 10 [94]. Cf. also 2 [86], 18 [100] and 33 [113]. 139 See n.123 above. 140 As noted by Runia (1997a) 95.



views of the mathematical astronomers. In addition, the commentators of Epicurus and Lucretius have traced the ultimate sources of many more of the alternative explanations. It appears that almost every one of their alternative explanations has been borrowed, sometimes with minor modifications, from one or other of their predecessors. Epicurus’ dependence on earlier theories was already recognised in antiquity. In doxographical reports Epicurus’ opinion, if included, is usually mentioned last,141 and expressed in terms which relate it to the preceding views. For instance, in Aët. II 13.15 (on the substance of the stars) we read: ÉEp¤kourow oÈd¢n épogin≈skei toÊtvn §xÒmenow toË §ndexom°nou.142

Epicurus rejects none of these (explanations), clinging to what is possible.

And in Aët. II 22.4 (on the shape of the sun): ÉEp¤kourow §nd°xesyai tå proeirhm°na pãnta.

Epicurus holds all the (explanations) to be possible.


A similar report is found in Seneca’s Naturales Quaestiones VI 20.5 (on earthquakes): Omnes istas esse posse causas Epicurus ait pluresque alias temptat.

Epicurus says that all these causes may apply and he tries his hand at several more.

Judging from these testimonies one gets the impression that Epicurus himself must have had before him some doxographical work, very much like Aëtius’ Placita, where he would have found all the relevant theories on each topic neatly listed side by side, which he could have simply copied out, striking the names of the original authors, expressing his consent with all of them, and sometimes adding a few of his own. That Epicurus might have followed such a procedure was first suggested by Diels and Usener, and has

141 142

Runia (1992) 135, n.76. For the formula “clinging to what is possible”, cp. Epicurus himself in Pyth. 9 [93]: pãnta går tå toiaËta ka‹ tå toÊtoiw suggen∞ oÈyen‹ t«n §narghmãtvn diafvne›, §ãn tiw ée‹ §p‹ t«n toioÊtvn mer«n §xÒmenow toË dunatoË efiw tÚ sÊmfvnon to›w fainom°noiw ßkaston toÊtvn dÊnhtai §pãgein. [“For all these and kindred explanations are not at variance with any clear-seen facts, if in such departments of inquiry one always clings to what is possible and can refer each point to what is in agreement with the appearances.” (tr. Bailey, slightly modified)]



been generally accepted since.143 For chronological reasons Aëtius’ Placita – the only virtually complete doxographical work that has come down to us – cannot itself have been Epicurus’ source (as is obvious from the inclusion of Epicurus’ name in Aëtius’ work), but it is believed that Aëtius’ work derives from earlier works of a similar nature, having a similar structure and lay-out, which may have been used by Epicurus.144 There are two important arguments in favour of this theory. In the first place it is clear that Lucretius’ representation of earlier views does sometimes depend on doxographical reports, rather than autopsy of the original works. This has been demonstrated convincingly by Wolfgang Rösler (1973) with respect to DRN I 635-920, where Lucretius successively deals with the views of Heraclitus, Empedocles and Anaxagoras concerning the ultimate constituents of reality. As Rösler pointed out, certain misrepresentations, generalisations and choice of terminology, such as the designation of the Anaxagorean first principles as homoeomeria,145 betray Lucretius’ dependence on a doxographical tradition that goes back to Aristotle and Theophrastus. In addition, David Sedley has pointed out that Lucretius’ critical survey of these three doctrines, including the word homoeomeria, may well derive from books XIV and XV of Epicurus’ On nature,146 which suggests that it was Epicurus himself who derived his knowledge of these Presocratic theories from doxography,147 and passed it on to Lucretius. This does not prove, of course, that Epicurus’ and Lucretius’ accounts of astronomical and meteorological theories, in the Letter to Pythocles and DRN V 509-770 and VI, depend on


Diels (1879) 225: “Epicuri epistula ad Pythoclem {…} tanquam ex doxographis nominibus philosophorum omissis raptim corrasa est {…}.” Usener (1887) xl-xli: “Elegisse autem Epicurum perquisitis omnium physiologorum libris quis credat? Quem etsi Democriti et Democriteorum, Anaxagorae et Archelai opiniones facile concedemus ipsum ex illorum libris novisse, reliquorum ut cognosceret rationes consentaneum est librum ei ad manum fuisse, quo conpositas et conparatas nullo negotio inveniret, hoc est Theophrasti fusik«n dÒjaw.” See also Ernout-Robin (1925-28), III 201-2; and Runia (1997a). 144 In Usener’s and Diels’ wake this earlier work, from which Aëtius’ Placita is supposed to be ultimately derived, is often identified with Theophrastus’ Fusika‹ dÒjai (‘Physical Opinions’), a work of which only a single fragment remains (see Runia (1992) 117). However, following the important studies on these subjects by Jaap Mansfeld and David Runia it seems more prudent now to simply state that Aëtius’ Placita is based on and influenced by, several works of Theophrastus as well as Aristotle. See e.g. Mansfeld (1989) esp. 338-42; Mansfeld (1992b) and Mansfeld (2005). 145 Lucr. DRN I 830 & 835; on the provenance of this term see Rösler (1973), 67-68 with nn. thereto, and Sedley (1998a), 124-125. 146 Sedley (1998a) 123-6 & 145-6. 147 As Rösler (1973) 71-72 already suspected.



doxographical reports as well, but it certainly adds to the likelyhood of this hypothesis. Another argument in favour of an ultimately doxographical origin of these passages is the close correspondence in the order of subjects that can be observed in – on the one hand – the meteorological sections of Lucretius’ DRN and Epicurus’ Letter to Pythocles, and – on the other hand – book III of Aëtius’ doxography.148 The question of the precise relations between these three texts is, however, complicated by the existence of a further parallel, a meteorological treatise ascribed to Theophrastus and preserved in Syriac and Arabic, which exhibits more or less the same order of subjects. We will examine this treatise and the nature of the complications involved more closely in §1.5.5 below, while the correspondences in the order and scope of subjects of these four works, as well as a number of other texts, will be investigated more thoroughly in Chapter Two. It is remarkable, however, that the observed correspondence in the sequence of subjects does not extend to the astromical sections of these works – here Epicurus, Lucretius and Aëtius differ considerably from each other both in the range and the order of subjects – nor to the order of the indivual explanations per subject – in this respect too Epicurus, Lucretius, Aëtius and, in addition, the Syriac meteorology differ considerably from each other.149 One aspect of Epicurean multiple explanations does not seem to be accounted for by the assumption of a doxographical origin. As we saw above, analogy plays an important role in the validation of individual explanations. Accordingly, in the DRN many explanations of astronomical and especially meteorological phenomena are supported by references to similar appearances here with us. For instance, in DRN VI 121-31 one possible cause of thunder – wind being trapped in a hollow cloud and then violently bursting forth – is compared to the explosion of an air-filled bladder. As Epicurus’ Letter to Pythocles and the corresponding books of Aëtius’ Placita (book II, on cosmology and astronomy, and book III, on meteorology and terrestrial 148

As observed by Reitzenstein (1924) 34-5; Runia (1997a) 97; Sedley (1998a) 158. See also Table 2-4 on p.117 below, where the order of subject of Lucretius VI, Epicurus Pyth., Aët. III and the Syriac meteorology are compared. 149 As observed by Sedley (1998a) 182. For instance, Lucretius’ ‘habit’ of mentioning the views of the mathematical astronomers first (see §1.3.6 on p.39ff above) is quite unlike the way Aëtius structures his chapters according to diaeresis and diaphonia. Aëtius’ method is becoming increasingly clear thanks to the works of David Runia and Jaap Mansfeld: see esp. Runia (1989) and (1992), and Mansfeld & Runia (2009a) part 1, pp.3-16 et passim, and part 2, passim.



phenomena) provide very few specific analogies, one might be tempted to ascribe the addition of such analogies to Lucretius himself rather than – through Epicurus’ mediation – a doxographical source. As we have seen, however, the virtual absence of specific analogies from Epicurus’ Letter to Pythocles may well be due to its being a summary of a more extensive work.150 In fact, many of the analogies provided by Lucretius are old, much older even than Epicurus – the comparison of thunder with an exploding bladder, for instance, is found in Aristophanes’ Clouds (lines 403-7) – and it seems unlikely that the theories and the accompanying analogies should have reached Lucretius by different roads. It also noteworthy that the Syriac meteorology, which I mentioned above, offers many of the same analogies as Lucretius, including the exploding bladder (1.18-20). In sum, it is very likely that the meteorological and astronomical portions of Lucretius’ DRN ultimately derive – probably for the most part through Epicurus’ works – from a doxographical source, which in the sequence of its subjects may have resembled Aëtius, but which, in contrast with Aëtius, combined the naked doxai with explanative analogies. What place should be assigned to the Syriac meteorology in this transmission is as yet unclear.

1.5 The sources of the method of multiple explanations 1.5.1 Introduction Epicurus and Lucretius may have derived most of their alternative explanations from doxography, but the result is something new and different. Stripped of their name-labels, theories devised by earlier thinkers have been transformed into truly alternative explanations endorsed by Epicurus and Lucretius themselves.151 So, even though doxography may explain where the individual explanations came from, it does not explain how Epicurus and Lucretius came to use them as they did. In this section I want to examine a number of other texts and authors who also sometimes resorted to the use of multiple explanations, and find out if and to what extent they resemble and may have influenced Epicurus. The examination will include Democritus, Aristotle, Theophrastus, and the Syriac meteorology commonly ascribed to Theophrastus. In order to be able to make a comparison, it will be useful first to indicate some general characteristics of the method as employed by Epicurus and Lucretius. The following features I consider typical of Epicurean multiple explanations: 150 151

See p.32 with notes 86 and 87 above. This is not always appreciated by the commentators: see Ernout-Robin (1925-28) 202: “c’est en somme une doxographie, mais sans nom propre.”



1. They are applied to non-evident matters, such as the nature and causes of astronomical and meteorological phenomena. 2. In those fields of physical enquiry where multiple explanations are used (astronomy and meteorology) they are used systematically. 3. Lists of multiple explanations may consist of up to eight or nine explanations.152 4. Many explanations, especially in meteorology, are supported with analogies from everyday experience. 5. Most alternative explanations can be related to the views of earlier thinkers. Armed with these five distinctive features we may now proceed to investigate possible parallels to Epicurus’ and Lucretius’ method of multiple explanations.

1.5.2 Democritus Already in antiquity Epicurean physics was often viewed as a modernised version of the teachings of Democritus.153 It is not unreasonable therefore to start our investigation into the origins of Epicurus’ method of multiple explanations with Democritus. A very promising testimony in this respect is provided by Seneca. In NQ VI 20 (Democritus fr.A98 D-K), having just presented an overview of ancient theories on earthquakes, Seneca continues thus: [1]

Veniamus nunc ad eos qui omnia ista quae rettuli in causa esse dixerunt aut ex his plura. Democritus plura putat. Ait enim motum (i) aliquando spiritu fieri, (ii) aliquando aqua, (iii) aliquando utroque, et id hoc modo prosequitur: …

Let us now come to those who said that all these causes, which I recounted, are responsible or several of these. Democritus thinks several. For he says that an earthquake (i) sometimes happens because of wind, (ii) sometimes of water, (iii) sometimes of both, and he pursues this in the following manner: …

Seneca goes on to describe each of these three explanations in some detail – first explanation (ii) and then (iii) and (i) – after which he continues, picking up the reference at the beginning of the chapter to ‘those who said that all these causes … are responsible’:


Epic. Pyth. 19 [101-2] offers eight causes of lightning, and DRN VI 96-159 nine causes of thunder. 153 Cf. Cic. N.D. I 73 “quid est in physicis Epicuri non a Democrito?” See also Cic. Acad. Post. I 6; De fin. I 17-18, 21; II 102; IV 13; and N.D. I 93 & 120; Diog. Laërt. X 4; Plut. Adv. Col. 3, 1108e-f.


Omnes istas esse posse causas Epicurus ait pluresque alias temptat ...


Epicurus says that all these causes may apply and he tries his hand at several more …

According to Seneca, then, both Democritus and Epicurus explained earthquakes with a number of alternative theories, but whereas Epicurus accepted all available theories, Democritus was more selective. This explicit contrast seems to suggest that Seneca knew something about Democritus’ method and how this differed from Epicurus’. It is quite probable, however, that Democritus’ selectivity is only apparent. Democritus wrote at a time when many of the theories later described by Seneca had not yet been devised and the major doxographical works reporting them not yet been written.154 I am inclined to think, therefore, that Seneca had no positive information about Democritus’ methodology at all, but simply inferred so much from the three alternative explanations he found attributed to Democritus, which he himself contrasted to the larger number of explanations offered by Epicurus. The amount of detail with which Seneca is able to report Democritus’ three explanations may suggest that he had a very good source for them, but in fact the wording of the text indicates that he may have filled in much of the detail by himself. At the beginning of section 2, for instance, he writes: ‘And now, just as we spoke of wind, we must also speak of water’,155 as if he were describing his own theory, instead of someone else’s. Seneca’s report on Democritus also strangely deviates from the account offered by Aristotle in Mete. II 7, 365b1-6 (Dem. fr.A97 D-K). According to Aristotle, Democritus held that earthquakes occur both (a) when the earth is full with water and its cavities overflow, and (b) when the earth is dried up and its cavities draw water from elsewhere. Aristotle’s first explanation may perhaps be identified with Seneca’s second, but Aristotle’s second explanation has nothing to do with either Seneca’s first or his third theory. In sum, it is quite possible that Democritus offered more than one explanation for earthquakes, but in view of the discrepancy between Aristotle’s and Seneca’s accounts we cannot be certain which explanations were his. Yet, neither Aristotle nor Seneca tell us why Democritus resorted to multiple explanations. For all we know Democritus may have simply offered his two or three explanations by way of a hypothesis, without any strong epistemological motives. There is no indication, moreover, that Democritus extended this use of multiple explanations to other phenomena as well. On the 154

Jaap Mansfeld points out to me that the fifth century BC already knew some doxographical overviews, such as Hippias’ presentation of related views and Herodotus’ account of the Nile flood, although these overviews fall far short of the doxographical works and passages of later times. See also Mansfeld & Runia (2009) 154ff. 155 Sen. NQ VI 20,2: “Etiam nunc, quomodo de spiritu dicebabus, de aqua quoque dicendum est.”



contrary, although there are many ancient reports concerning Democritus’ views on specific astronomical, meteorological and terrestrial problems, none of these (beside those on earthquakes) attribute to Democritus anything other than single explanations. There is no good reason therefore to consider Democritus a major source of inspiration for Epicurus’ method of multiple explanations.156

1.5.3 Aristotle A more promising candidate in this respect is Aristotle.157 Several cases of multiple explanations are found in his works, especially in the Meteorology. Here, in I 3, 341a12-31, Aristotle gives two reasons why the sun, though not fiery in itself, produces heat on earth. In I 4, 341b36 - 342a13, he offers two explanations for the production of shooting stars and the like. Finally, in I 7, 344a5-b4 he gives two accounts of the production of comets, corresponding to two different types of this phenomenon. This last subject is introduced with the following lines (344a5-8): ÉEpe‹ d¢ per‹ t«n éfan«n tª afisyÆsei nom¤zomen flkan«w épodede›xyai katå tÚn lÒgon, §ån efiw tÚ dunatÚn énagãgvmen, ¶k te t«n nËn fainom°nvn Ípolãboi tiw ín œde per‹ toÊtvn mãlista sumba¤nein.

For we consider that we have given a sufficiently rational explanation of things non-evident to senseperception if we have referred them to what is possible; and, on the basis of the present appearances, one may assume that they are best accounted for as follows.

Here, just like Epicurus later, Aristotle applies multiple explanations to things non-evident, inferring the possibility of each explanation on the basis of the appearances.158 It must be observed, though, that while Aristotle is here thinking of the appearances of the object of inquiry itself, Epicurus usually refers to the appearances of analogous phenomena here with us. Sometimes Aristotle too supports his alternative explanations by reference to appropriate analogies, as we can observe in his first account of the sun’s heat-production (341a23-27):


Pace Asmis (1984) 328-9. Asmis (1984) 329-30, Mansfeld (1994) 33, n.18. 158 Cf. Epic. Pyth. 9 [93]: pãnta går tå toiaËta ka‹ tå toÊtoiw suggen∞ oÈyen‹ t«n §narghmãtvn diafvne›, §ãn tiw ée‹ §p‹ t«n toioÊtvn mer«n §xÒmenow toË dunatoË efiw tÚ sÊmfvnon to›w fainom°noiw ßkaston toÊtvn dÊnhtai §pãgein. [“For all these and kindred explanations are not at variance with any clear-seen facts, if one always clings in such departments of inquiry to what is possible and can refer each point to what is in agreement with the appearances.” (tr. Bailey, slightly modified)] 157

MULTIPLE EXPLANATIONS TÚ d¢ mçllon g¤gnesyai ëma t“ ≤l¤ƒ aÈt“ tØn yermÒthta eÎlogon, lambãnontaw tÚ ˜moion §k t«n par' ≤m›n gignom°nvn: ka‹ går §ntaËya t«n b¤& ferom°nvn ı plhsiãzvn éØr mãlista g¤gnetai yermÒw.


That the heat is increased by the presence of the sun is easily enough explained by considering analogies from our own experience: for here too the air in the neighbourhood of a projectile becomes hottest. (tr. Lee)

Yet, in spite of these similarities there is still a huge gap between Epicurus’ and Aristotle’s approaches to multiple explanations. In the first place, Aristotle only very rarely recurs to multiple explanations: in the entire body of the Meteorology, only three clear cases are found. Most often Aristotle is perfectly happy to give just one explanation. Secondly, the number of alternatives offered is much smaller: in the Meteorology Aristotle in each case offers no more than two explanations, whereas Epicurus and Lucretius may offer up to eight or even nine possible causes. Thirdly, Aristotle only occasionally uses analogies to support an explanation, whereas for Lucretius and Epicurus analogy with everyday phenomena is essential for accepting an explanation. Finally, Aristotle’s multiple explanations do not seem to relate to earlier views: his accounts of the sun’s heat-production, of shooting stars and of comets proceed from Aristotle’s own physical theory. When, on the other hand, Aristotle does engage with older theories, he usually rejects them, and substitutes them with a single theory of his own making.159

1.5.4 Theophrastus With Theophrastus, Aristotle’s successor as head of the Lyceum, the use of multiple explanations becomes much more prominent. Many instances are found in Theophrastus’ minor treatises De ventis, De lapidibus and De igne, and many more in his botanical writings, especially his De causis plantarum.160 Yet, even in these works multiple explanations, though by no means rare, are still the exception. When offering multiple explanations, Theophrastus most often gives two, but occasionally more; the maximum number I have found is five.161 Sometimes the explanations offered are explicitly linked to analogous occurrences within our sensory experience, as can be seen in De igne 1, 4-11:


So Taub (2003) 94: “In the Meteorology, Aristotle does not usually accept the views of his predecessors, even when they are those of ‘the majority or the wise’.” 160 Steinmetz (1964) 32-33, 46, 82, 88, 91, 103, 122-123, 132, 139; Eichholz (1965) 6; Wöhrle (1985) 145-8; Vallance (1988) 34-5; Daiber (1992) 285; Gottschalk (1998) 287. 161 CP I 17.5.



ÖEti d¢ afl gen°seiw aÈtoË {sc. toË purÚw} afl ple›stai [ka‹] oÂon metå b¤aw, ka‹ går ≤ plhgª t«n stere«n Àsper l¤yvn, ka‹ tr¤cei ka‹ pilÆsei kayãper t«n pure¤vn ˜sa ¶xei forãw, Àsper t«n puroum°nvn ka‹ thkom°nvn (§k d' aÈtoË toË é°row ka‹ to›w n°fesi sustrofa‹ ka‹ yl¤ceiw: b¤aioi går dØ afl fora¤, di' œn dØ ofl prhst∞rew ka‹ kerauno‹ g¤nontai), ka‹ ˜souw dØ trÒpouw êllouw teyevrÆkamen e‡y' Íp¢r g∞w ginom°nvn e‡t' §p‹ g∞w e‡y' ÍpÚ g∞w. afl går polla‹ dÒjeian ín aÈt«n metå b¤aw.

Moreover, most forms of generation of fire take place by force, as it were; for instance, that caused by the striking of solids like stones, and those caused by friction and compression, as in firesticks and in all those substances which are in process, such as those which are ignited and fused (in fact, it is from air that the clouds undergo their concentrations and compressions, for of course the motions by which firewinds and thunderbolts are generated are forcible), and whatever other ways we have observed, whether above the earth, on it, or beneath it. Most of these appear to come about by force. (transl. Coutant)

Theophrastus only rarely comments on his motives for accepting several explanations; the only clear instance I have found is De lapidibus I 3, 1-3: ÑH d¢ p∞jiw to›w m¢n épÚ yermoË to›w d' épÚ cuxroË g¤netai. kvlÊei går ‡svw oÈd¢n ¶nia g°nh l¤yvn Íf' •kat°rvn sun¤stasyai toÊtvn.

This solidification is due in some cases to heat and in others to cold, for there may be nothing to prevent certain kinds of stone from being formed either by heat or by cold. (transl. Eichholz).

It is interesting to note that – apart from the modest ‡svw (‘perhaps’) – Theophrastus’ «kvlÊei går oÈd¢n» (‘for nothing prevents’) is very similar to the formulas Epicurus later uses to signal the validity of his alternative explanations, like «oÈd¢n går éntimarture›» (‘for nothing contests’) and «oÈy¢n §mpodostate›» (‘nothing stands in the way’).162 It must also be observed that, when Theophrastus offers multiple explanations, these hardly ever relate to earlier views, and when he does adduce older theories it is usually to reject them and replace them with a theory of his own.163 Until now we may have been comparing apples and oranges: Epicurus and Lucretius applied multiple explanations, not to fire or stones or different winds, but to astronomy and meteorology, and the few instances of multiple explanations in Aristotle’s works also occur in his Meteorology. It would be interesting therefore to see to what extent Theophrastus used multiple explanations in his meteorological treatise. The Syriac meteorology which is commonly, but in my view prematurely, ascribed to Theophrastus will be dealt with in the next subsection. Here I will confine myself to Greek and Latin testimonies of Theophrastus’ meteorological views. A very interesting 162 163

See p.18 n.49 above. See e.g. Theophr. De igne 52-56; HP III 1, 4-5; III 2, 2. Cp. what was said about Aristotle in n.159 above and text thereto.



text in this respect is fr.211B FHS&G, preserved by Olympiodorus In Arist. Mete. I 9, 346b30 (p.80.30-81.1 Stüve): ÉIst°on d°, ˜ti m¢n ı ÉAristot°lhw a‡tion l°gei t∞w efiw Ïdvr metabol∞w tØn cÊjin mÒnon. YeÒfrastow d¢ oÈ mÒnon tØn cÊjin afit¤an fhs‹ t∞w toË Ïdatow gen°sevw, éllå ka‹ tØn p¤lhsin. fidoÁ går §n Afiyiop¤& mØ oÎshw cÊjevw ˜mvw ÍetÚw katãgetai diå tØn p¤lhsin: fhs‹ går ˆrh e‰nai ¶keise ÍchlÒtata, efiw ì tå n°fh prospta¤ousi ka‹ e‰yÉ oÏtvw ÍetÚw katarrÆgnutai diå tØn ginom°nhn p¤lhsin. éllå mØn ka‹ §p‹ t«n lebÆtvn ÍgrÒthw, fhs¤n, éntikatãrrei, §t‹ d¢ ka‹ §p‹ t«n yÒlvn t«n loutr«n mØ paroÊshw cÊjevw, diå tØn p¤lhsin dhlonÒti toÊtou ginom°nou.

One should known that Aristotle says that cooling alone causes the transformation into water. Theophrastus, however, says that not only cooling causes the generation of water, but also condensation. Note that in Aethiopia where there is no cooling, rain nevertheless pours down due to condensation. For he says there are very high mountains there, against which the clouds collide and that subsequently rain pours down because of the ensuing condensation. Yet also in the case of cauldrons, says he, moisture runs down, and also in the case of the vaults of baths, where there is no cooling, this obviously occurs due to condensation.

According to this report,164 Theophrastus accepted two different causes of rain-production: cooling and condensation. The first view is authorized by Aristotle, for the second one example (rain in Aethiopia) and two supporting analogies (cauldrons and the vaults of baths) are adduced. Both views can be related to earlier theories: the first view is Aristotle’s (Mete. I 9, 346b30-31), as Olympiodorus informs us, and the second corresponds to the views of several earlier thinkers.165 For rain production, then, Theophrastus accepted two explanations. There is one other meteorological problem for which Theophrastus’ view is explicitly reported. In NQ VI 13.1 (fr.195 FHS&G), Seneca ascribes to Theophrastus and Aristotle the single view that earthquakes come about through exhalations rising from the earth and then, for lack of place, turning back on themselves. If this report is correct, Theophrastus did not use multiple explanations to account for every meteorological problem. Our findings may be summarized as follows. In his meteorological treatise (to judge from Greek and Latin testimonies), as in his works on fire, stones, winds and plants, Theophrastus frequently uses multiple explanations, although single explanations remain the norm. When he does give several explanations, the number of alternatives rarely exceeds two, although cases 164

Cf. Proclus In Plat. Tim. 22E (= Theophr. fr.211A FHS&G); Galen In Hippocr. Aer. 8.6 (= fr.211C ibid.), and Theophrastus De ventis 5.1-5. 165 Anaximenes A17 D-K (= Aët. III 4.1), Xenophanes A46 D-K (= Aët. III 4.4), Hippocrates Aer. 8.7 (II 34 Littré) and Democritus apud Diod. I 39.3 (cf. fr.A99 D-K = Aët. IV 1.4) all ascribe rain formation to condensation of clouds; this condensation being due, according to Hippocrates, to compression by contrary winds and other clouds, and, according to Democritus, to the clouds being compressed against high mountains.



with up to five can be found. Alternative explanations are occasionally supported with analogies from everyday experience, but more often they are not. They may be derived from the views of earlier thinkers, but in general they are not. In short, most aspects of Epicurean multiple explanations can be found in Theophrastus, but on a far more modest scale.

1.5.5 The Syriac meteorology In the previous subsection I have deliberately left out of account a treatise which most scholars now agree in identifying with (a part or a summary of) Theophrastus’ lost meteorological treatise, the Metarsiology.166 I have done so because I think there is still reasonable doubt about this identification. Since, however, this treatise furnishes the closest existing parallel to Epicurus’ and Lucretius’ use of multiple explanations, it must here be dealt with. The treatise is preserved in three versions: two mutually independent Arabic translations of a Syriac original, and a single, badly mutilated, copy of this Syriac original.167 In both Arabic versions the work is ascribed to Theophrastus (the corresponding section of the Syriac manuscript is lost). Together, the three versions seem to provide a good basis for a reconstruction of (the contents of) the Syriac original.168 The treatise covers a range of meteorological phenomena,169 explaining most of them with a number of alternative explanations. This pervasive use of multiple explanations in a treatise purportedly written by Theophrastus caused some embarrassment to the earlier commentators. They could not believe that Theophrastus would have employed a method so radically different (as it appeared to them) from Aristotle’s and Theophrastus’ other works.170 In fact, as Gotthelf Bergsträßer observed,171 if the treasise had 166

E.g. Daiber (1992) 282ff; Mansfeld (1992a) 314-16; id. (1994) 30; Sedley (1998a) 158, 179; Sharples (1998) 17, 144; Taub (2003) 116. 167 The Syriac version was first edited, with German translation and commentary, by Wagner & Steinmetz (1964). The Arabic version of Bar Bahlūl was first edited, with a German translation and commentary, by Bergsträßer (1918). The second Arabic version, probably made by Ibn Al-Khammār, was edited, with an English translation and commentary by Daiber (1992), who also offered improved editions of the other two versions, unfortunately without translation. 168 Not necessarily of the Greek original, as Daiber (1992) claims on pp.219 & 282-3. 169 The subjects are: (1) thunder, (2) lightning, (3) thunder without lightning, (4) lightning without thunder, (5) why lightning precedes thunder, (6) thunderbolts, (7) clouds, (8) rain, (9) snow, (10) hail, (11) dew, (12) hoar-frost, (13) winds (including an account of the prhstÆr), (14) the halo around the moon, and (15) earthquakes. 170 See e.g. Strohm (1937) 411: “daß die verworrene Folge von sieben Erklärungen des Donners und vier Gründen des Blitzes, die in einer an Epikurs Probabilismus



not been explicitly ascribed to Theophrastus, no one would have guessed that it was his. In view of its offering multiple explanations for most of the phenomena, and in view of the close parallels with DRN VI, both in the order of subjects (on which see pp.115ff below) and in the treatment of individual subjects, attribution to Epicurus or his school would have seemed obvious. For this reason Bergsträßer and Boll, and later Reitzenstein and a few others, felt compelled at least to consider the possibility of an Epicurean origin.172 Unfortunately, they did not follow up on this hypothesis, but simply dismissed it in favour of another hypothesis, to the effect that the treatise’s Greek original had been “a doxography or at least a discussion with a strong doxographical character [...], in which the excerptor had deleted the names of the inventors of each individual theory as being of no consequence, and rather blurred the traces of the author’s own position.” 173 In this way they were able to uphold Theophrastus’ authorship, which has never been seriously doubted again,174 even though the views about the treatise’s real character have radically changed. Since Daiber’s publication of the text with translation and commentary in 1992, it is generally accepted that the treatise really offers multiple alternative explanations, and not, as was previously believed, a (critical) doxography.175 Yet, if this is true (as I believe it is), the grounds on which Bergsträßer and Reitzenstein were able to reject the possibility of an Epicurean rather than a Theophrastean origin of the treatise, seem to have been removed: Theophrastus’ authorship can no longer be taken for granted. I will therefore refer to the treatise simply as the ‘Syriac meteorology’ and from this neutral position try to establish how the treatise relates to Theophrastus on the one hand and Epicurus and Lucretius on the other. Although, as we saw above, Theophrastus was not as averse to multiple explanations as the earliest commentators of the treatise believed, there is still gemahnenden Reihe nebeneinanderstehen, in dieser Form Theophrast fremd ist, liegt auf der Hand.” 171 Bergsträßer (1918) 28. 172 Bergsträßer (1918) 28; Boll in his epilogue to Bergsträßer (1918) 30; Reitzenstein (1924) 7-11; Drossaert Lulofs (1955) 438. The question is wisely left open by Robin in Ernout-Robin (1925-28) III 200-1 and 249. 173 Bergsträßer (1918) 28. See also Reitzenstein (1924) 8ff, Strohm (1937) 411; WagnerSteinmetz (1964) 14, 34; Steinmetz (1964) 55. 174 Bergsträßer (1918), 28, also advances the possibility of a late compendium, and so does Gottschalk (1965), 759-60, who identifies some passages as deriving from Strato rather than Theophrastus. Kidd (1992), 294, also leaves open the possibility of a late compendium, and Van Raalte (2003) believes that the so-called Theological Excursus cannot derive from Theophrastus. Doubts about the correctness of the attribution to Theophrastus also in Ernout-Robin III 200-1. 175 Daiber (1992) 285; Kidd (1992) 303-4; Mansfeld (1992a) 325; Mansfeld (1994) 33; Sedley (1998a) 181; Sharples (1998) 17 with n.58; Taub (2003) 116-7.



a huge gap between the use of multiple explanations in the Syriac meteorology and in Theophrastus’ undisputed writings: 1. While in Theophrastus’ undisputed writings single explanations still seem to be the rule, in the Syriac meteorology most problems are explained in a number of ways.176 2. While in Theophrastus’ undisputed writings the number of alternative explanations rarely exceeds two, the Syriac meteorology frequently offers three or more, the maximum number being seven (as in the case of thunder).177 3. While in Theophrastus’ undisputed writings alternative explanations are only rarely supported with analogies, the Syriac meteorology abounds with them.178 4. While in Theophrastus’ undisputed writings alternative explanations only rarely derive from earlier, Presocratic, theories, in the Syriac meteorology the vast majority of theories can be identified with specific views of earlier thinkers.179 In all these respects the Syriac meteorology is much closer to the meteorological sections of Epicurus’ Letter to Pythocles and Lucretius’ De rerum natura than to Theophrastus’ undisputed works. It has been observed that in the Syriac meteorology different explanations sometimes apply to different types of a certain phenomenon, each exemplified by a different analogy.180 It has also been suggested that in this respect the use of multiple explanations in the Syriac meteorology differs from the way they are used by Epicurus and Lucretius, who typically conceived of their several explanations as equally possible alternative causes of a single undifferentiated phenomenon.181 However, this supposed contrast between the Syriac meteorology and Epicurus / Lucretius is based on a misrepresentation of the latter: as we have seen above (p.11), in his meteorology Lucretius too occasionally differentiates phenomena by type, and Epicurus probably did so as well. With respect to meteorological phenomena, therefore, there is no 176

See APPENDIX C on p.246 below: 26 cases of multiple explanations vs. 22 with single explanations. 177 See APPENDIX C on p.246 below: 9 cases with three or more explanations vs. 17 with just two explanations. 178 See Daiber (1992) 284, 285, 288 (‘illustrative examples’ or ‘illustrative experiments’) and esp. Taub (2003) 117-20. 179 Daiber (1992) 287-88, 290; Taub (2003) 117. 180 Daiber (1992) 279 (ad 13.22-32), 285, 288; Kidd (1992) 299-304; Sharples (1998) xv; Taub (2003) 117, 130-131; Garani (2007) 97. 181 Kidd (1992) 303-4; Sharples (1998) xv; Taub (2003) 130-131; Garani (2007) 97.



observable difference between the application of multiple explanations in the Syriac meteorology and in the meteorological sections of Epicurus and Lucretius. The real difference must be looked for outside the scope of meteorology. Although the Syriac meteorology only deals with meteorological phenomena (which traditionally included earthquakes), the treatise does contain an oblique reference to another branch of physical enquiry. In 14.14-17 we read the following: Neither the thunderbolt nor anything that has been mentioned [i.e. other meteorological phenomena] has its origin in God. For it is not correct (to say) that God should be the cause of disorder in the world; nay, (He is) the cause of its arrangement and order. And that is why we ascribe its arrangement and order to God (...) and the disorder of the world to the nature of the world. (transl. Daiber)

Whereas Epicurus and Lucretius do not acknowledge any basic distinction between (disorderly) meteorological and (orderly) astronomical phenomena, neither of which are the work of God, the Syriac meteorology leaves God in charge of everything orderly in the world, which probably refers to the orderly arrangement of the world as a whole, as well as the orderly motions of the heavenly bodies. This also settles part of the question about the Syriac meteorology’s authorship: the work as we have it, or at least the theological remark just quoted, could not have been written by Epicurus himself, although the work as a whole may still be influenced by him. The question of the treatise’s origin and relations with Epicurus and Lucretius will be considered more thoroughly in Chapter Two (p.132ff) below.

1.5.6 Conclusions concerning the origins of multiple explanations It is now time to formulate some conclusions concerning the possible origins of Epicurus’ method of multiple explanations. However great Democritus’ influence on Epicurean physics may have been, there is little reason to assume that he was a major source of inspiration for Epicurean multiple explanations as well. Even if he gave two or three alternative explanations of earthquakes, there is nothing to suggest that this was the inevitable result of certain epistemological considerations, or that he extended the use of multiple explanations to other physical problems as well. The first clear instances of multiple explanations are found in the (undisputed) writings of Aristotle and Theophrastus. Although in these works multiple explanations are not applied systematically, whenever they are, they seem to be confined to a certain class of phenomena. Aristotle remarks that



non-evident things need to be accounted for with reference to what seems possible on the basis of the present appearances – a procedure which, apparently, will sometimes result in the acceptance of multiple explanations –, and Theophrastus ascertains the validity of his alternative explanations from the observation that nothing prevents any one of them. Aristotle and Theophrastus (in his undisputed works) rarely derive their alternative explanations from doxography, and when they do deal with earlier views they usually refute them. It is very likely that Epicurus’ use of multiple explanations was somehow inspired by Aristotle and Theophrastus. Yet, the precise relationship between Epicurus and his two predecessors depends on our position with respect to the authorship of the Syriac meteorology. If we assume that the Syriac meteorology is, as is commonly believed, either the whole or a part or a summary of Theophrastus’ lost Metarsiology, we may conclude as follows. For some reason Theophrastus has decided that those epistemological conditions which invite the acceptance of multiple explanations and which occasionally obtain in other fields of physical inquiry, must be operative in meteorology most of the time. He must further have decided that alternative explanations in meteorology can be drawn from the works of earlier philosophers, whose views in other fields of inquiry he usually rejects. Epicurus’ own special contribution would have been to extend the use of multiple explanations from meteorology to cosmology and astronomy (using a doxographical work, perhaps Theophrastus’ Physical opinions,182 as a source for individual explanations), and to provide the method with a rigorous epistemological justification. If, on the other hand, the Syriac meteorology is assumed to somehow presuppose Epicurus’ work and therefore postdate it, Epicurus’ contribution must have been much more substantial. Perhaps taking his inspiration from the scattered instances of multiple explanations in the works of Aristotle and Theophrastus, and from the accompanying epistemological remarks, he himself fashioned them into a veritable method with a rigorous epistemological basis, and he himself decided that doxography might provide the necessary alternative explanations.

1.6 Conclusions In this chapter various aspects of Epicurus’ and Lucretius’ method of multiple explanations have been explored. Section 1.2 provides a brief 182

On Theophrastus’ Physical opinions see §2.5.5 on p.141 below.



overview of the various uses of multiple explanations. Section 1.3 discusses some epistemological aspects of the method. Section 1.4 presents the main arguments for an ultimately doxographical origin of Epicurus’ and Lucretius’ alternative explanations, and section 1.5 discusses a number of ancient parallels for the use of multiple explanations outside Epicureanism. The most important findings of this chapter are the following. In §1.3.2 three modern theories concerning the truth-value of multiple explanations have been investigated. One theory, according to which alternative explanations are only collectively true, could be rejected on account of the evidence. Yet, it appeared to be impossible to choose between the two remaining theories. Some of the evidence suggests that Epicurus claimed that all alternative explanations are true, while another part of the evidence favours the view that alternative explanations can at best be called possible. In §1.3.5 it has been argued that Diogenes of Oenoanda’s claim that some explanations are more plausible than others is a departure from Epicurus and Lucretius for whom all alternative explanations have the same truth-value. In §1.3.6 Bailey’s assertion that Lucretius, by his habit of mentioning the views of the mathematical astronomers first, betrays his preference for their views, has been refuted, and a polemical motive discovered instead. Finally, in §1.5, I have argued that Aristotle’s and Theophrastus’ occasional use of multiple explanations and their epistemological justifications for this use may have inspired Epicurus’ method of multiple explanations. Theophrastus’ influence on Epicurus may turn out to be even greater if the Syriac meteorology can be proved to be Theophrastus’. However, comparison of the use of multiple explanations in Theophrastus’ undisputed works and the Syriac meteorology indicates that Theophrastus’ authorship of the latter work is still far from certain. The question of the Syriac meteorology’s authorship will be further explored in the next chapter.


2.1 Introduction The sixth book of Lucretius’ DRN and Epicurus’ Letter to Pythocles are often described as meteorological treatises.183 I am not going to contend that they are not, but it is remarkable how ill-defined the word meteorology actually is, and how variable its subject matter. The Letter to Pythocles, whose subject matter is described as ta meteōra, ‘the things above’, discusses both atmospherical and astronomical phenomena and earthquakes. Lucretius’ book VI, on the other hand, leaves out astronomy altogether but includes a number of terrestrial phenomena, such as the size of the sea, eruptions of Etna, the summer flooding of the Nile, poisonous places, curious wells and springs, magnets and diseases. The reason that both works are sometimes called ‘meteorologies’ seems to be that both of them give considerable space to those phenomena which we still designate as meteorological: i.e. weather phenomena. This characteristic they share with several other ancient writings, in particular Aristotle’s Meteorology, which gave this branch of physical inquiry its name, and which is in fact the only ancient meteorological work to use this word. Other notable examples are Seneca’s Naturales Quaestiones (‘Natural Questions’) and the Syriac meteorology ascribed to Theophrastus. Each of these works deals with a different selection of phenomena, but the core of their subject matter is meteorological in the modern sense of the word, and therefore they can be called meteorologies. However, although the precise boundaries of the subject matter may vary between one meteorology and the next, they are by no means arbitrary, but depend on certain underlying assumptions and traditions, which may be brought to light by a thorough comparison of the extant ancient meteorologies. Such a comparison could also shed light on the various ways in which the subject matter is sometimes subdivided.In this chapter I propose to carry out such a comparison of ancient meteorological writings, in order to elucidate the various traditions and the position of Lucretius and Epicurus therein. Special attention will be given to the second part of Lucretius’ book VI, which is largely devoted to exceptional local phenomena or mirabilia. Although such phenomena are mentioned in other meteorologies as well, they belong more properly to paradoxography. I will therefore extend my investigation to this genre as well, in order to define the precise relations and 183

See e.g. Ernout-Robin III 199-200; Taub (2003) 127-37.


the division of labour between the two genres, and Lucretius’ position vis-àvis the two. Another matter is the order in which the various meteorological subjects are discussed. It has often been observed that the order of subjects in Lucretius’ book VI closely resembles those of the Syriac meteorology and Aëtius’ book III, and to a lesser degree Epicurus’ Letter to Pythocles, but the precise extent of these similarities and the exceptions to them have never been thoroughly assessed. In this chapter I will delve into this matter as well, with a view to establishing the relations between these four works. In this context it will also be necessary to deal with the question of the identity of the Syriac meteorology, which is generally – but in my view prematurely – identified with Theophrastus’ Metarsiologica, but could in fact be a later work based largely on Epicurus’ meteorology, which would turn the tables between the two works. The structure of this chapter will be as follows. In §2.2, I will compare a number of ancient meteorological writings with respect to the range, delimitation and subdivision of their subject matter. Then, in §2.3, the second part of Lucretius’ DRN VI, which is devoted to the explanation of predominantly exceptional local phenomena, will be compared to other meteorological as well as paradoxographical works. In §2.4 the order of subjects of Lucretius’ DRN VI, Epicurus’ Letter to Pythocles, the Syriac meteorology, and book III of Aëtius’ Placita will be compared and an original order proposed. In §2.5 the question of the Syriac meteorology’s identity will be investigated, and possible relations between the four texts indicated. Finally, in §2.6, the major conclusions of the chapter will be summarized.

2.2 Range, delimitation and subdivisions of meteorology 2.2.1 Introduction In this section I will compare a number of writings dealing exclusively or for the most part with meteorology. By meteorology I mean the study of atmospherical phenomena as well as such phenomena as were often associated with them, such as the Milky way, comets, shooting stars, earthquakes and terrestrial waters. The comparison in this section will be confined to such matters as the range of subjects of meteorology, its delimitation from astronomy and its major subdivisions. I will not, in this section, go into the the treatment of individual subjects or the theoretical background of each work,184 unless these throw some light upon the reasons for including a certain subject in meteorology or in one of its major subdivisions. From this point of view I 184

For individual subjects see the commentaries to the relevant passages. For ancient meteorology in general see Taub (2003).



am only interested in those works or testimonies that present a reasonably complete and coherent account of meteorology, and I will pass by the meteorological views, however interesting they may be, of e.g. Posidonius or Arrianus, which are known to us from scattered references only.185 The writings that meet the above requirements are the following186 (the page numbers indicate the beginning of my brief introduction of each work): • • • • • • • • •

Aristotle Meteorology, books I-III [Aristotle] De mundo, ch. 4 Aëtius Placita, book III (+ IV 1) Pliny Naturalis Historia, book II, §§89-248 Seneca Naturales Quaestiones Stoics apud Diog. Laërt. VII 151-4 The Syriac meteorology Epicurus Letter to Pythocles, 16ff [98ff]187 Lucretius De rerum natura, book VI

p.73 p.76 p.78 p.81 p.82 p.82 p.84 p.85 p.88

2.2.2 Ancient meteorological texts Before I set out on my task, it will be expedient to offer something of a definition of meteorology, to state what it is, how it is delimited from other fields of investigation, notably astronomy, and how (if at all) it is subdivided. The English word ‘meteorology’ derives from the Greek metevrolog¤a, which is the study of tå met°vra, the ‘lofty’ things. Before Aristotle the word met°vrow and its derivatives appear to have been used indiscriminately to refer to both astronomical and atmospherical phenomena.188 This does not necessarily mean that philosophers before Aristotle did not somehow distinguish between these two fields of physical inquiry.189 That at least Democritus did, is suggested by the presence of two separate titles, Afit¤ai 185

Posidonius’ meteorological fragments are collected in Edelstein & Kidd (1972) (frs.11, 15, 121 and 129-38; see also frs.214-29 on tides and hydrology and frs.12 and 230-2 on seismology) and Arrian’s in Roos & Wirth (1967/8), vol. 2, pp.186-195. 186 All these works but one (viz. Aëtius III) are discussed in Taub (2003), and all but one (viz. Pliny NH II) feature in the appendix ‘On the Order of Presentation of Meteorological Phenomena’ in Kidd (1992) 305-6. 187 Passages in the Letter to Pythocles will be referred to by the chapter number of Bollack & Laks (1978), with the traditional numbering of Meibom’s edition of Diogenes Laërtius in square brackets. 188 Capelle (1912a) 421-41; id. (1935), col.316. 189 As Capelle seems to think: see Capelle (1912a) 425, 427, 447-8; id. (1913) 322; id. (1935), col.316, lines 15-16, 22-28.


oÈrãniai (celestial causes) and Afit¤ai é°rioi (atmospherical causes), in the catologue of his writings in Diog. Laërt. IX 47. Yet, it is to Aristotle that we owe the first clear delimitation of the two fields with respect to each other, and the restriction of the terms met°vrow and metevrolog¤a to sublunary phenomena only. Interestingly, in antiquity Aristotle’s limitation of meteorology to the sublunary sphere was more influential than the name he gave to it: two generations later Epicurus could still use the word met°vrow to refer to both meteorological and astronomical phenomena, and the famous Stoic scholar Posidonius too dealt with cosmological and astronomical matters under the single heading of meteorology,190 while Aristotle’s associate and successor Theophrastus found it necessary to introduce another word altogether (metãrsiow and metarsiolog¤a) to refer to atmospherical phenomena.191 Aristotle Meteorology In the opening chapter of his Meteorology192 Aristotle defines the province of meteorology as ‘everything which happens naturally, but with a regularity less than that of the first element of material things, and which takes place in the region which borders most nearly on the movements of the stars’.193 Further down he says: ‘The whole region around the earth, then, is composed of these bodies {i.e. earth, water, air and fire}, and it is the conditions which affect them which, we have said, are the subject of our inquiry’.194 While thus delimiting meteorology from astronomy, at the same time Aristotle extends its subject matter to include not just atmospherical, but also fiery, watery and even earthly phenomena. Astronomy, which is the subject of his De Caelo, deals with the orderly and eternal movements of the heavens and the stars, 190

Posidonius is reported to have defined ‘cosmos’ in a work called MetevrologikØ stoixe¤vsiw (D.L. VII 138 = fr.14 E-K), to have discussed the substance of the sun in his Per‹ mete≈rvn (D.L. VII 144 = fr.17 E-K), and contrasted the physical and mathematical approaches to astronomy in his Metevrologikã (Simpl. In phys. 291.21292.21 = fr.18 E-K), all of which are either cosmological or astronomical matters. On the other hand he is reported to have discussed the rainbow in a work called MetevrologikÆ (D.L. VII 152 = fr.15 E-K). See also Capelle (1913) 337ff. 191 See Capelle (1913) 333-6. 192 For a general account of Aristotle’s Meteorology see Taub (2003) 77-115. 193 Arist. Mete. I 1, 338b1-3, ˜sa sumba¤nei katå fÊsin m°n, étaktot°ran m°ntoi t∞w toË pr≈tou stoixe¤ou t«n svmãtvn, per‹ tÚn geitni«nta ... mãlista tÒpon tª forò tª t«n êstrvn. Transl. Lee (1952), slightly modified. According to Lee (1952), xii note a, this refers to the entire sublunary region, according to Capelle (1912b), 516-517, to its fiery upper part only. 194 Arist. Mete. I 2, 339a19-21, ı dØ per‹ tØn g∞n ˜low kÒsmow §k toÊtvn sun°sthke t«n svmãtvn: per‹ o tå sumba¤nonta pãyh fam¢n e‰nai lhpt°on. Transl. Lee (1952), slightly modified.



which are made up of the ‘first element’ (a.k.a. aether); meteorology on the other hand studies the less orderly phenomena of the region around the earth, which is occupied by the four classic elements, earth, water, air and fire. The boundary between the two regions is marked by the orbit of the moon. Yet, Aristotle’s definition is not extremely precise and leaves open many questions. It is instructive therefore to have a closer look at the actual range of phenomena covered by Aristotle’s Meteorology. I will leave book IV out of account, as it is generally agreed that it was not part of the original work.195 In books I-III, then, the following subjects are discussed (I have Italicised those subjects we would nowadays no longer call meteorological)196: A. PHENOMENA OF THE FIERY UPPER ATMOSPHERE Shooting stars I4 Other luminary phenomena I5 Comets I 6-7 The Milky way I8 B. PRODUCTS OF MOIST EXHALATION Mist, clouds and rain I9 Dew and hoar-frost I 10 Snow I 11 Hail I 12 C. PRODUCTS OF DRY EXHALATION (PLUS TERRESTRIAL WATERS) Winds I I 13 Springs and rivers I 13  Climatic and coastal change I 14  TERRESTRIAL WATERS The sea II 1-3  Winds II II 4-6 Earthquakes II 7-8 Thunder and lightning II 9 Thunderbolts and whirlwinds III 1 D. NON-SUBSTANTIAL PHENOMENA Rainbows I III 2 Haloes III 2-3 Rainbows II III 4-5 Rods and mock suns III 6 E. PHENOMENA OF THE EARTH Minerals and metals III 6


As was observed already by Alexander of Aphrodisias (2nd cent. AD) In Arist. Mete. 4.1, 179.1-5 Hayduck. See refs. to modern literature in Taub (2003) 206 n.24. 196 For the organization of Meteorology I-III see e.g. Capelle (1912b) or Louis (1982) XXVIII- XXXIV.


As the table shows, Aristotle’s Meteorology covers more than just atmospherical phenomena. In the first place, it includes a number of subjects which we would nowadays call astronomical, like the Milky Way, comets and shooting stars, probably owing to their – real or apparent – irregularity, which to Aristotle seemed incompatible with the supposed orderly character of the supralunary world.197 Although not meteorological in our sense of the word, these phenomena are at any rate met°vra, i.e. ‘lofty things’. This is not true of some other phenomena included in Aristotle’s Meteorology, viz. rivers and the sea, earthquakes, and minerals and metals. Minerals and metals seem to be included because they too, just like some of the phenomena above the earth, are products of the two exhalations.198 Earthquakes are more closely connected with the met°vra proper. They are treated right after winds, because they are themselves caused by subterranean winds.199 Not so clear are Aristotle’s motives for including rivers and the sea. As the table shows, the entire passage on rivers and the sea is inserted in the section on winds. This move seems to have been prompted by certain analogies between wind and flowing water. Such analogies, however, do not in themselves justify the classification of rivers and the sea as meteorological phenomena. A more plausible reason for their inclusion in a work on meteorology would be that rivers and the sea, being sustained by rain and melting snow, are an integral part of the hydrological cycle,200 and thus intimately related to the subject of meteorology, but Aristotle does not adduce this justification. Whatever Aristotle’s reasons were, from then on rivers and the sea were frequently included in works on meteorology. Aristotle’s Meteorology does not deal with volcanoes as a separate subject, although it occasionally refers to volcanic phenomena as by-products of earthquakes.201 Neither does it concern itself with problems pertaining to the 197

Already in antiquity Aristotle was rightly criticised for assigning the Milky Way to meteorology. See e.g. Olymp. In Arist. Mete. [CAG 12.2] 10.33; 66.17-20; 75.24-76.5 (citing Ammonius); Philop. In Arist. Mete. [CAG 14.1] 113.33-118.26 (citing Damascius 116.36ff). It may therefore not be a coincidence that the Milky Way is absent from almost all subsequent writings on meteorology (Aëtius excepted). Based on Theophr. fr.166 FHS&G, Steinmetz (1964) 167-8 concludes that Aristotle’s view was already rejected by Theophrastus; for a more critical attitude concerning the evidence see Sharples (1985) 584-5 and id. (1998) 108-111. 198 Arist. Mete. III 6, 378a13 ff. 199 Arist. Mete. II 7, 365a14-15 (quoted on p.94 below). Cf. ibid. II 8, 366a3-5, oÈk ín oÔn Ïdvr oÈd¢ g∞ a‡tion e‡h, éllå pneËma t∞w kinÆsevw {sc. t∞w g∞w}, ˜tan e‡sv tÊx˙ =u¢n tÚ ¶jv énayumi≈menon. – “So the cause of an earthquake is likely to be neither water nor earth but wind, when the external exhalation happens to flow inwards”. 200 On the hydrological cycle see Arist. Mete. I 9, 346b22-347a8; I 13, 349b3-8; II 2, 354b28-34 and II 3, 356b22-357a2. 201 See e.g. Arist. Mete. II 8, 367a1-11 on the eruption of the Aeolian island of Hiera.



earth as a whole, like its shape, position and stability, which had already been treated in Aristotle’s De Caelo.202 [Aristotle] De mundo 4 Practically the same subject matter is also discussed in chapter 4 of the De mundo.203 Although the work is clearly rooted in Aristotelian philosophy, most scholars reject Aristotle’s authorship and date the work to some time between 50 BC and 150 AD.204 The subject of chapter 4 is introduced as ‘the most notable phenomena in and about the inhabited world (i.e. land and sea)’.205 The structure of the chapter can be set out as follows:


Arist. Cael. II 13-14. The occasional references to such matters in the Meteorology (I 3, 339b7-9 on the relatively small size of the earth; I 9, 346b24 on the earth being at rest; II 5, 362a33-b33 on the five terrestrial zones) merely serve as a background for real meteorological problems. 203 See Taub (2003) 161-8. 204 See Furley (1955) 337-341. Mainly on linguistic and stylistic grounds G. Reale en A.P. Bos believe that the work’s ascription to Aristotle is correct (Reale (1974); Reale & Bos (1995); etc.), while J. Barnes (1977) and D.M. Schenkeveld (1991), though excluding Aristotle’s authorship, argue for a much earlier date than has hitherto been accepted. Their arguments have failed to convince the majority of scholars: see esp. J. Mansfeld (1991), 541-3, and (1992c). Although the question of the work’s authorship is no concern of the present work, it may yet contribute to an answer: it will be shown (see Table 2-1 below) that the De mundo differs from Aristotle’s Meteorology in its subdivision of the subject matter, its omission of the Milky Way (see also n.197 above), and its inclusion of tides, volcanoes and poisonous exhalations (on the inclusion of the two last subjects in meteorology see p.107 below). 205 394a7-8: Per‹ d¢ t«n éjiologvtãtvn §n aÈtª {sc. gª ka‹ yalãtt˙, ¥ntina kale›n efi≈yamen ofikoum°nhn} ka‹ per‹ aÈtØn pay«n nËn l°gvmen.

EPICURUS & LUCRETIUS AND THE SCOPE AND STRUCTURE OF ANCIENT METEOROLOGY 77 GENERAL INTRODUCTION A. PHENOMENA OF THE AIR 1. The two exhalations 2. Products of the moist exhalation (mist, dew, ice, hoar-frost, dew-frost, cloud, rain, snow, hail) 3. Products of the dry exhalation (winds, violent winds, incl.: thunder, lightning, thunderbolt, prhstÆr, tuf≈n) 4. Appearances vs. substantial {sc. luminary} phenomena 5. Appearances (rainbow, ‘rod’, halo) 6. Substantial {sc. luminary} phenomena (s°law, shooting star, comet) 206 B. PHENOMENA IN THE EARTH (hot springs, volcanoes, noxious exhalations, earthquakes) C. PHENOMENA IN THE SEA (chasms, retreats and incursions of waves, submarine volcanoes, springs and rivers, trees growing in the sea (!), currents, eddies, tides) GENERAL CONCLUSION

394a7-8 394a9 – 395b17 394a9-19 394a19 – b6 394b7 – 395a28 395a28-32 395a32 – b3 395b3-17 395b18 – 396a16 396a17-27


As the table shows the subject matter is basically divided into three parts: phenomena of the air, phenomena in the earth, and phenomena in the sea.207 Unlike Aristotle’s Meteorology, the De mundo seems to recognize volcanoes as phenomena to be studied in their own right, mentioning them (incl. the Etna) briefly but separately before a longish discussion of earthquakes. Like 206

In De mundo 2, 392a32-b5 these same phenomena are referred to the layer of fire above the air, not to the air itself. 207 In 395b17-18 we read: Tå m¢n to¤nun é°ria toiaËta. ÉEmperi°xei d¢ ka‹ ≤ g∞ pollåw §n aÍtª, and in 396a17 it says: Tå d¢ énãlogon sump¤ptei toÊtoiw {sc. to›w §n gª pãyesi} ka‹ §n yalãss˙ ... Finally, in the general conclusion (396a27-32), we read: ÑVw d¢ tÚ pçn efipe›n, t«n stoixe¤vn §gkekram°nvn éllÆloiw §n é°ri te ka‹ gª ka‹ yalãss˙ katå tÚ efikÚw afl t«n pay«n ımoiÒthtew sun¤stantai, to›w m¢n §p‹ m°rouw fyoråw ka‹ gen°seiw f°rousai, tÚ d¢ sÊmpan én≈leyrÒn te ka‹ ég°nhton fulãttousai. By the way, in chapters 2 and 3 the same subject matter is organized differently. There, having first dealt with a number of cosmological and astronomical issues (391b9 – 392a32), the author moves on to the phenomena of the sublunary sphere. This falls apart into three regions: that of fire (392a32-b5), that of air (392b5-13), and that of earth and sea taken together (392b14 – 394 a6). To the fiery region are attributed such phenomena as comets and shooting stars. The air is the abode of clouds, rain, snow, frost, hail, winds, whirlwinds (tuf«new), thunder, lightning and thunderbolts. In the subsequent section on the earth and the sea, the author, rather than enumerate the corresponding physical phenomena, offers a picturesque geographical description of the terrestrial sphere, which need not concern us here.



Aristotle’s Meteorology the meteorological section of the De mundo excludes problems concerning the earth as a whole. Aëtius Placita III More or less the same subject matter is dealt with in the third book of Aëtius’ Placita.208 This work, to be dated most likely to the first century AD, has not come down to us directly, but can be reconstructed to a certain degree from two later works that largely derive from it: Pseudo-Plutarch’s Placita and Stobaeus’ Eclogae Physicae. Of these two Pseudo-Plutarch has most faithfully preserved the work’s original division into books and chapters,209 and it is this division which is generally followed and which we shall follow too. Book III of the Placita is not a meteorology in the sense of the two works mentioned above. While Aristotle’s Meteorology and the De mundo aim to give a coherent theory of the whole field of meteorology, Aëtius is concerned with presenting the various and often conflicting views brought forward by earlier thinkers. Yet the subject matter of the book coincides largely with, and betrays a strong dependence on, Aristotelian meteorology in the range, subdivision and order of its subjects.210 The book lacks a single general heading: having dealt with cosmology and astronomy in the previous book, in book III Aëtius goes on to discuss, first, in chapters 1-8, what he calls tå metãrsia (lofty phenomena),211 and then, in chapters 9 and following, what he calls tå prÒsgeia (down-to-earth phenomena).212 The latter section also 208

On Aëtius’ work in general see now Mansfeld & Runia (1997) and id. (2009a). On book III in particular see Mansfeld (2005). 209 Diels (1879) 61; Mansfeld & Runia (1997) 184-5. 210 On Aristotle as a source for Aëtius as to methodology and contents see Mansfeld (1992b). On Aëtius III depending on, and deriving from, Arist. Mete. I-III see Mansfeld (2005). See also Mansfeld & Runia (2009a). 211 Aëtius III 0: Perivdeuk∆w §n to›w prot°roiw §n §pitomª tÚn per‹ t«n oÈran¤vn lÒgon, selÆnh dÉ aÈtvn tÚ meyÒrion, tr°comai §n t“ tr¤tƒ prÚw tå metãrsia: taËta dÉ §st‹ tå épÚ toË kÊklou t∞w selÆnhw kayÆkonta m°xri prÚw tØn y°sin t∞w g∞w, ¥ntina k°ntrou tãjin §p°xein tª perioxª t∞w sfa¤raw nenom¤kasin. ÖArjomai dÉ §nteËyen. – “Having briefly traversed in the previous chapters the account of the heavenly phenomena, of which the moon is the border region, I shall in the third book turn to lofty phenomena. These are what is from the circle of the moon to where the earth is situated, which they are convinced occupies the position of the centre in relation to the circumference of the sphere. I shall begin from here.” On this passage see now Mansfeld & Runia (2009a) 54, whose translation I have basically followed. 212 Aëtius III 8.2: Perigegramm°nvn d° moi t«n metars¤vn, §fodeuyÆsetai ka‹ tå prÒsgeia. – “The lofty phenomena having been described by me, the down-to-earth


includes a number of chapters, 9-14, dealing with the earth as a whole,213 which are not part of the scope of Aristotle’s Meteorology or chapter 4 of the De mundo. Their inclusion may have been prompted by Aëtius’ wish to present his subjects in a rigorous top-down (or ‘outside-in’) order, based on the location of each cosmic part and each phenomenon rather than its nature. The structure of the book (as preserved by Pseudo-Plutarch) is as follows: TA METARSIA 1. milky way 2. comets, shooting stars and the like 3. thunder, lightning, thunderbolts and whirlwinds (typhōnes and prēstēres) 4. clouds, rain, snow and hail 5. rainbow 6. rods and mock suns 7. winds 8. winter and summer TA PROSGEIA 9. the earth (being unique and limited) 10. shape of the earth 11. position of the earth 12. inclination of the earth 13. motion of the earth 14. division of the earth 15. earthquakes 16. the sea: its origin and bitterness 17. the sea: ebb and flood 18.* halo IV 1.* the flooding of the Nile (Stob. 39* water properties)


Two chapters appear to have been misplaced. Chapter 18 on the halo does not belong in the section on tå prÒsgeia, which it now concludes, but must have been part of the preceding section on tå metãrsia. There is in fact quite some evidence to connect it more specifically with chapters 5 and 6, on the rainbow and ‘rods and mock suns’ respectively.214 Also misplaced is the first chapter of book IV, which discusses the topic of the Nile flood. This subject phenomena, too, will be inspected.” On this passage see now Mansfeld & Runia (2009a) 55. 213 On Aëtius’ inclusion of these subjects with meteorology see p.93 below. 214 On the dislocation of Aët. III 18 see Diels (1879) 56, 60-61, Lachenaud (1993) 25, Mansfeld (2005) 26-27, 37 (n.52) and 56 and Mansfeld & Runia (2009a) 44. Note that in Arist. Mete. III 2-6, [Arist.] De mundo 4, 395a32-b3, and Sen. NQ I 2-13, the subjects of the rainbow, the halo and rods and mock suns are also discussed successively (see also p.121 below).



was clearly meant to go with tå prÒsgeia in the second part of book III, where also the origin and salinity of the sea and the causes of ebb and flood are dealt with.215 It is quite out of place in book IV, which is otherwise about the soul and its functions. It is possible that a further chapter existed, which is now missing. In Stobaeus’ Eclogae Physicae, one of the two main sources for the reconstruction of Aëtius’ text,216 there is a chapter (39), titled Per‹ Ídãtvn (‘On waters’),217 which has no counterpart in Pseudo-Plutarch’s Placita, the other main source for Aëtius and our principal guide as to the table of contents of Aëtius’ work.218 In its present state the chapter contains only one lemma reporting Aristotle’s views on water properties, which probably derives not from Aëtius but from Arius Didymus (fragment 14a).219 However, since the subjects covered by Stobaeus’ Eclogae Physicae derive to a large extent from Aëtius220 (even though Stobaeus often adds or substitutes lemmas from other sources), it is possible that this chapter’s title and subject too derive from Aëtius. That the chapter in its present state contains no Aëtian material can be ascribed to the very selective transmission of that part of Stobaeus’ work which corresponds to the second half of Aëtius book III.221 Stobaeus’ chapter ‘On waters’ follows immediately upon the chapter on tides and so has the same relative position as Aëtius’ chapter on the flooding of the Nile, which only Pseudo-Plutarch has preserved. If, as I have suggested, Stobaeus’ chapter ‘On waters’ derives from Aëtius, it will have immediately preceded, or followed on, the chapter on the Nile flood. It is worth noting at this point that in Seneca’s Naturales Quaestiones (see below) a book dealing with 215

See Diels (1879) 56 and 61, Lachenaud (1993) 274. On Stobaeus as a source for the reconstruction of Aëtius, see Mansfeld & Runia (1997) 196-271. 217 Not included in the list of possibly lost chapters in Mansfeld & Runia (1997) 186, but see the table printed ibid. pp.214-6. 218 On Pseudo-Plutarch as a source for the reconstruction of Aëtius, see Mansfeld & Runia (1997) 121-195. 219 Diels (1879) 854; see also Mansfeld & Runia (1997) 249 n.167. 220 Mansfeld & Runia (1997) 216: “the topics covered by the book [i.e. Stobaeus’ Eclogae Physicae] have been largely based on the subjects dealt with in the Placita [of Aëtius]. Only 7 or 8 of the 60 chapters find no equivalent in A[ëtius].” 221 Mansfeld & Runia (1997) 202-3: “When we further examine the epitomized chapters ¶31-60 in Book I, we soon observe that a very one-sided selection has taken place. Only lemmata containing Platonica, Aristotelica, Pythagorica and Hermetica are retained. In various chapters that must have contained copious extracts from Aëtius just one or two lemmata containing the views of Plato and Aristotle are written out (¶32, 36, 38-39, 4243, 45, 51-60).” (my emphasis). 216


(terrestrial) waters (III) is followed by a book on the summer flooding of the Nile (IVa). Pliny Naturalis Historia II §§ 89-248 Roughly the same range of subjects, including (like Aëtius) a number of sections dealing with the earth as a whole,222 is discussed in the second part of book II of Pliny’s Naturalis Historia.223 The first part of the book is devoted to cosmology and astronomy. Pliny devides astronomical and meteorological phenomena differently from Aristotle, Pseudo-Aristotle and Aëtius. Not only erratic celestial phenomena like comets and shooting stars, but also merely apparent phenomena like haloes, ‘rods’ and mock suns (but not the rainbow!) are classified by Pliny among the ‘stars’. The text of book II (like every other book) abounds in repetitions, interruptions and all kinds of excursuses, which make it hard to summarize its contents. The following overview (from §89 onwards) is no more than an impression. For a more complete summary see Pliny’s own table of contents in book I of the Naturalis Historia. A. Cosmos/heavens and stars (i-xxxvii, §§1-101) xxii-xxxvii §§ 89-101 ‘Sudden stars’ (incl. comets, shooting stars, but also haloes and other insubstantial luminary phenomena) B. Atmospherical phenomena (xxxviii-lxii, §§102-153)224 xxxviii §§ 102-4 Nature of air xxxix-xli §§ 105-10 Influence of astronomy on the weather and animals xlii § 111 Rain, wind and clouds xliii §§ 112-3 Storm-winds and thunderstorms xliv-l §§ 114-134 Winds (incl. whirlwinds) li-lvi §§ 135-146 Thunderbolts lvii-lix §§ 147-150a Miraculous phenomena in and from the sky lx §§ 150b-151 Rainbows lxi § 152 Hail, snow, hoar-frost, mists, dew, clouds lxii § 153 Particular local climates.


On Pliny’s inclusion of these subjects with meteorology see p.93 below. On Pliny’s meteorology see Taub (2003) 179-187. 224 At the beginning of §102, referring back to the previous sections, including those on comets, shooting stars, haloes and ‘rods’, Pliny writes: “Hactenus de mundo ipso sideribusque. Nunc reliqua caeli memorabilia: namque et hoc caelum appellavere maiores quod alio nomine aëra.” 223


CHAPTER TWO C. Earthly phenomena (lxiii-cxiii, §§154-248)225 lxiii-lxxx §§ 154-190 The earth as a whole (shape, position, seasons) lxxxi- lxxxvi §§ 191-200 Earthquakes lxxxvii - xciv §§ 201-6a Formation of new land xcv §§ 206b-208 Products of the earth (incl. mines, gems, peculiar stones, medicinal springs, volcanoes, poisonous exhalations) xcvi § 209 Vibrating lands and floating islands xcvii-xcviii §§ 210-1 Local earth marvels xcix-civ §§ 212-23 Tides (and other effects of the moon and sun) cv § 224a Depth of the sea cvi §§ 224b-234a Miraculous waters and some universal properties of springs cvii-cix §§ 234b-235 fiery phenomena cx §§ 236-8 volcanoes cxi §§ 239-241 marvels of fire cxii-cxiii §§ 242-248 size of the earth

In Table 2-1 below (p.91) I will not include every oneof the many subjects touched upon by Pliny, but only those which have a clear counterpart in one or more of the other texts. Seneca Naturales Quaestiones More or less the same range of subjects is also covered by Seneca’s Naturales Quaestiones.226 At the outset of the second book,227 Seneca divides the study of natural phenomena into three parts: the caelestia (heavenly things), the sublimia (‘lofty’ things) and the terrena (earthly things). The term caelestia denotes the phenomena of the heavens and the heavenly bodies, i.e. cosmology and astronomy. The sublimia cover all phenomena occurring in the region between the heavens and the earth, i.e. atmospherical phenomena, but also earthquakes. Finally, among terrena are understood such subjects as waters, lands, trees, plants and ‘everything contained in the ground’.228 If we 225

The last words of §153 are: “Haec sint dicta de aëre.” §154 starts with: “Sequitur terra.” See Taub (2003) 141-161. On the macro-structure of the NQ as compared to Seneca’s own programmatic remarks at NQ II 1, 1-2 see Mansfeld & Runia (2009a) 46-48 and 119-121. 227 According to Carmen Codoñer Merino (1979), xii-xxi, and independently Hine (1981), 6-19, originally the eighth and last book. 228 Sen. NQ II 1, 1-2: “Omnis de uniuerso quaestio in caelestia, sublimia, terrena diuiditur. Prima pars naturam siderum scrutatur et magnitudinem et formam ignium quibus mundus includitur, solidumne sit caelum ac firmae concretaeque materiae an ex subtili tenuique nexum, agatur an agat, et infra sese sidera habeat an in contextu sui fixa, quemadmodum anni uices seruet, solem retro flectat, cetera deinceps his similia. | 226


compare the range of subjects actually covered in the Naturales Quaestiones with Seneca’s theoretical division of natural phenomena, it appears that the work is concerned almost exclusively with sublimia and terrena. There is only one exception: comets, which according to Seneca should be classed with the caelestia.229 However, in holding this view he is, as he himself admits, dissenting from the accepted Stoic (and Aristotelian) view that comets are irregular and therefore necessarily sublunary phenomena.230 By including the subject in a treatise otherwise devoted to atmospherical and earthly phenomena, Seneca is simply following the tradition. In the Naturales Quaestiones Seneca does not deal with questions concerning the earth as a whole, but he informs us that some of these (esp. those concerning the earth’s position) should be classified among the caelestia,231 while the rest belong with the terrena. Two parts of the work appear to be lost: book IVa, on the Nile flood, lacks its final part, and IVb, presently on hail and snow, its beginning. It is likely that IVb originally included such subjects as clouds and rain as well.232 The work as it has come down to us covers the following subjects: I II III IVa IVb V VI VII

Lights in the sky (both substantial an insubstantial) Lightnings and thunders Terrestrial waters (almost entirely excluding the sea) Nile Hail and snow Winds, incl. whirlwinds Earthquakes Comets

The overview of Stoic meteorology in Diogenes Laërtius VII 151-4 In chapters 151-154 of book VII of his Lives of Eminent Philosophers, Diogenes Laërtius offers an overview of the Stoic theories on ‘things taking Secunda pars tractat inter caelum terramque uersantia. Hic sunt nubila, imbres, niues, et humanas motura tonitrua mentes; quaecumque aer facit patiturue, haec sublimia dicimus, quia editiora imis sunt. Tertia illa pars de aquis, terris, arbustis, satis quaerit et, ut iurisconsultorum uerbo utar, de omnibus quae solo continentur.” 229 Sen. NQ VII 22 1.1-3: “Ego nostris {sc. Stoicis} non assentior. Non enim existimo cometen subitaneum ignem sed inter aeterna opera naturae.” See also VII 4 and 21.1. 230 Comets are included unreservedly in Diogenes Laërtius’ overview of Stoic meteorology (VII 151-154), in Aristotle’s Meteorology (I 6-7), and in the meteorological sections of the De mundo (4, 395b8-9) and Aëtius (III 2). 231 Seneca, NQ II 1, 5: “ubi quaeretur quis terrae situs sit, qua parte mundi consederit, quomodo aduersus sidera caelumque posita sit, haec quaestio cedet superioribus et, ut ita dicam, meliorem condicionem sequetur.” 232 Corcoran (1971/2), ‘Introduction’ xx; Hine (1981) 10, 29-30; Gross (1989) 185.



place in the air’ (VII 151.1: t«n dÉ §n é°ri ginom°nvn).233 The overview comprises the following subjects (I have numbered them as they occur in the text): 1. Seasons 2. Winds 3. The rainbow 4. Comets 5. Shooting stars 6. Rain 7. Hoar-frost

8. Hail 9. Snow 10. Lightning 11. Thunder 12. Thunderbolts 13. Whirlwinds (typhōnes & prēstēres) 14. Earthquakes.

As Table 2-1 (on p.91 below) will show, this range of subjects corresponds almost exactly to that of Seneca’s sublimia (if we include comets). The Syriac meteorology In the previous chapter the Syriac meteorology has been introduced already (see §1.5.5 on p.64ff above). This work, which is preserved in one Syriac and two Arabic versions, is now commonly believed to be either the complete text of, or an extract from, Theophrastus’ lost two-book treatise Metarsiologikã.234 For reasons explained above I am not convinced this identification is certain, although the work is obviously Greek in origin. In the treatise as we have it the following subjects are discussed: 1. Thunder 2. Lightning 3. Thunder without lightning 4. Lightning without thunder 5. Why lightning precedes thunder 6. Thunderbolts 7. Clouds 8. Rain

9. Snow 10. Hail 11. Dew 12. Hoar-frost 13. Wind, incl whirlwinds (prēstēres) 14a. Halo around the moon 14b. Theological excursus 15. Earthquakes

There has been much speculation about whether or not the treatise as we have it might be complete. It has been suggested that it might originally have included subjects like comets and shooting stars, the rainbow and a number of terrestrial phenomena other than earthquakes. In this section I will try to avoid


On D.L.’s account of Stoic meteorology see Taub (2003) 137. On the origins of D.L.’s account of Stoic philosophy in general (VII 38-160) see Mansfeld (1986); Mejer (1978) 5-7; and id. (1992) 3579-82. 234 Cited by Diogenes Laërtius (V 46) in his list of works by Theophrastus.


such speculation; instead I will compare the text as we have it with other Graeco-Roman meteorologies and see where that may lead us. Epicurus Letter to Pythocles The Letter to Pythocles is one of Epicurus’ three doctrinal letters quoted in full in the tenth book of Diogenes Laërtius’ work on the lives and doctrines of the philosophers.235 In the introduction Epicurus claims that with this letter he is complying with Pythocles’ request for a ‘concise and well-described account of “lofty matters”’ (per‹ t«n mete≈rvn sÊntomon ka‹ eÈper¤grafon dialogismÚn). Pythocles had complained that what Epicurus had written elsewhere (tå går §n êlloiw ≤m›n gegramm°na) on these matters was hard to remember. The Letter’s subject, “lofty matters” (tå met°vra), is nowhere clearly defined, but appears to cover not just meteorological, but also cosmological and astronomical matters. Below I will provide a table of contents of the entire letter. The chapter numbers are those of Bollack & Laks (1978), which I think provide more insight into the structure of the text (in addition I have split chapters 17 and 27, which each deal with two separate subjects, into an A and a B part); the traditional numbering of Meibom’s edition of Diogenes Laërtius will be added in the second column. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.


Introduction Method Definition of ‘cosmos’ Number and origin of cosmoi Formation of the heavenly bodies Size of the heavenly bodies Risings and settings Motions of the heavenly bodies Turnings of the sun and moon The phases of the moon The light of the moon The face in the moon Eclipses of the sun and moon The heavenly bodies’ regular periods The length of nights and days

p.84-85 p.85-88 p.88 p.89-90 p.90-91 p.91 p.92 p.92-93 p.93 p.94 p.94-95 p.95-96 p.96-97 p.97 p.98

            


The authenticity of the Letter to Pythocles has been a matter of some doubt: Usener (1887) xxxvii-xli; Reitzenstein (1924) 36-43; Bailey (1926) 275; Schmidt (1990) 34-7. Bollack & Laks (1978) 45-55 provide a good overview and a convincing refutation of the arguments against the Letter’s authenticity. See also Mansfeld (1994) n.2 and Sedley (1998a) n.65.


CHAPTER TWO 16. 17A. 17B. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27A. 27B. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37.

Weathersigns Clouds Rain Thunder Lightning Why lightning precedes thunder Thunderbolts Whirlwinds (prēstēres) Earthquakes (Subterranean) winds Hail Snow Dew Hoar-frost Ice The rainbow The halo around the moon Comets Revolution of the stars Planets Lagging behind of certain stars Shooting stars Weathersigns from animals Conclusion

p.98-99 p.99 p.99-100 p.100 p.101-102 p.102-103 p.103-104 p.104-105 p.105-106 p.106 p.106-107 p.107-108 p.108 p.109 p.109 p.109-110 p.110-111 p.111 p.112 p.112-113 p.114 p.114-115 p.115-116 p.116

                → →    → →



Many scholars have commented upon the Letter’s strange order of subjects.236 At first the order is clear enough: chapters 1-2 are introductory, 35 deal with cosmological matters, 6-15 are astronomical, and 16-29 meteorological (including earthquakes). It is at this point that the confusion begins: some of the following chapters (32-34) deal with subjects that are undeniably astronomical again, while others (30, 31 and 35) are concerned with subjects that most ancient meteorologists considered meteorological, although Pliny classified them as astronomical (as Table 2-1 on p.91 below will show). Chapter 36 is again meteorological and seems in fact to be a supplement to what was said in chapter 16. Although Epicurus does not formally distinguish between astronomical and meteorological phenomena, both of which he simply calls “lofty matters” (met°vra), the structure of the letter suggests that he did recognize some kind 236

Usener (1887) xxxviii-xxxix; Reitzenstein (1924) 36, 40-3; Bailey (1926) 275; Arrighetti (1973) 524, 691ff; Bollack & Laks (1978) 11-18; Sedley (1998a) pp.122-3 with n.75, p.157.


of division. Unfortunately, the confused order at the end of the letter prevents us from establishing with certainty where Epicurus would have placed the dividing line: haloes (30), comets (31) and shooting stars (35) might all be either meteorological or astronomical. However, the way in which the sections on comets and shooting stars straddle three undeniably astronomical subjects seems to favour their classification as astronomical phenomena.237 A comparison with a number of other texts may help us to decide this matter (see Table 2-1 on p.91 below). In the introduction to the Letter Epicurus refers to what he had written elsewhere (tå går §n êlloiw ≤m›n gegramm°na) on these matters. We do not know where exactly Epicurus had dealt with these subjects before, but from fragments and citations we do know that at least some cosmological and astronomical problems were discussed in books XI and XII of his On nature. There is no evidence of any actually meteorological phenomena being discussed anywhere in the On nature, although it is clear that Epicurus must have discussed at least some of them outside the Letter to Pythocles as well, witness e.g. the long Epicurean account of earthquakes in Seneca’s Naturales Quaestiones VI 20.5-7, which cannot derive from the corresponding passage in the Letter to Pythocles. In his reconstruction of Epicurus’ On nature, David Sedley suggests that meteorological phenomena may have been discussed in book XIII of this work, separated from the discussion of astronomical phenomena in books XI and XII by the interposition of a passage on other worlds and the origin of civilisation, which would have occupied the later part of book XII. If this reconstruction is correct, this would suggest that Epicurus did indeed distinguish between astronomy and meteorology. However, as Sedley himself admits, this part of his reconstruction is highly speculative, and it cannot be excluded that meteorological phenomena were dealt with in book XII immediately following the discussion of astronomical phenomena.238 In the Letter to Pythocles Epicurus does not deal with subjects concerning the earth as a whole, but we know that at least one such subject, the earth’s stability, was discussed at the end of book XI of his On nature, following and preceding a number of astronomical subjects in books XI and XII.239 This suggests that, as far as Epicurus did distinguish between astronomical and meteorological phenomena, problems pertaining to the earth as a whole were classed with the former.


So Usener (1887) xxxviii. Sedley (1998a) 122-123 with note 76. 239 Sedley (1998a) 119-122. 238



Lucretius De rerum natura VI In book VI of the DRN Lucretius discusses a number of atmospherical and terrestrial phenomena roughly coinciding with the subjects of Aristotle’s Meteorology, and other works of this genre. More precisely, the following subjects are dealt with (or, in the case of snow, wind, hail, hoar-frost and ice, simply enumerated and then passed over): 96-159 160-218 219-422 423-450 451-494 495-523 524-526 527-534 535-607 608-638 639-702 703-711 712-737 738-839 840-847 848-878 879-905 906-1089 1090-1286

Thunder Lightning Thunderbolts Whirlwinds (prēstēres) Clouds Rain Rainbow Snow, wind, hail, hoar-frost, ice Earthquakes Why the sea does not grow bigger Etna The principle of multiple explanations The summer flooding of the Nile Avernian places Water in wells colder in summer The spring of Hammon A cold spring which kindles tow The magnet Diseases

Lucretius does not, in this context, deal with comets and shooting stars, nor does he discuss them in the astronomical passage in book V 509-770. Shooting stars (‘noctivagaeque faces caeli flammaeque volantes’) are mentioned in an overview of astronomical and meteorological phenomena in V 1189-93, but in such a way that it cannot be made out in which of the two groups Lucretius would have classed them. Yet, the evidence seems to be slightly in favour of Lucretius’ assigning comets and shooting stars to astronomy. The overview of atmospherical phenomena in book VI is quite exhaustive and even those subjects, like snow, wind, hail, hoar-frost and ice, which he chooses not to discuss, he still feels obliged to mention. Had he felt that comets and shooting stars belong to this class too, he would probably have mentioned them too. The account of astronomical phenomena in book V, on the other hand, is rather selective. Here Lucretius could have omitted comets and shooting stars without explicitly saying so. Problems pertaining to the earth as a whole are not discussed in the meteorological passage either. Lucretius does, however, discuss one such


problem – the stability of the earth – in book V (534-563), right in the middle of his astronomical section, which suggests that he might have considered other problems concerning the earth as a whole as belonging in that class as well. Book VI falls apart into two main divisions, the first dealing with atmospherical, and the second with terrestrial phenomena. There is some incertainty about the exact place of the cut, especially with respect to the section on earthquakes (535-607). In lines 527-534 (i.e. just before the account of earthquakes) Lucretius invites the reader to find out for himself the causes of “the other things that grow above and are produced above” (527, tr. RouseSmith), such as snow, wind, hail, hoar-frost and ice. This seems to imply that the account of these “things that grow above and are produced above” is hereby concluded, and that all subsequent subjects, beginning with earthquakes, belong to another class of phenomena.240 It is also possible, however, to place the dividing line after the subject of earthquakes. The next subject, the constant size of the sea, starts with the following words (608-9): Principio mare mirantur non reddere maius naturam, ...

In the first place people wonder why nature doesn’t make the sea bigger, ..

The word ‘principio’ (‘in the first place’) seems to suggest that Lucretius is now passing on to something new, viz. phenomena that inspire wonder –, of which the constant size of the sea presents the first instance.241 Below we shall further explore this group of problems, both in relation to the preceding section in Lucretius, and to more or less corresponding sections in other meteorological works (see §2.3 on p.99ff below). Comparison with other meteorological accounts may tell us, among other things, how each of the two proposed divisions relates to the traditional divisions of the subject, and to what extent Lucretius fits in this tradition.

2.2.3 The table Table 2-1 on the following page provides a synopsis of the subjects that are dealt with in each of the nine ‘meteorologies’. Subjects that are explicitly excluded from meteorology by the respective authors have been shaded grey. Also indicated are the major subdivisions of the subject matter as applied in


Bailey (1947), 1567 and 1632, classifies all phenomena up to and including snow, wind, hail, hoar-frost and ice (i.e. lines 96-534) with ‘atmospheric phenomena’ and the rest (i.e. lines 535-1137), including earthquakes, with ‘terrestrial phenomena. Cf. Giussani (1896-8), ad DRN VI 535-607. 241 So Giussani (1896-8), ad DRN VI 608-638, and Bailey (1947), 1646-7.



each of the nine texts. The order in which the texts and the subjects are presented is my own and is not here at stake.242 For the sake of brevity, passages of Pseudo-Aristotle’s De mundo 4 are indicated by the last digit of the Bekker page only, the first two digits of the relevant pages being always 39. In Diogenes Laërtius’ account of Stoic meteorology I have applied my own numbering according to the order in which the subjects are presented by Diogenes. Epicurus’ Letter to Pythocles is structured according to the chapter numbers of the edition of Bollack & Laks (1978), to which I have made the minor adjustment of dividing chapters 17 and 27 each into an A and a B part.


The order of subjects in a number of texts is dealt with in §2.4 on p.115 below.

III 2 & 4-5

II 9 II 9 III 1 III 1 I9 I9 I9 I 11 I 12 I 10 I 10


Thunder Lightning Thunderbolts Whirlwinds (prēstēres etc.) Mist Clouds Rain Snow Hail Dew Hoar-frost Ice Winds Seasons Weathersigns

Constant size of the sea Coastal change Origin & salinity of the sea Tides Minerals and metals Magnets Diseases

Volcanoes (Etna) Poisonous exhalations Springs and rivers Miraculous waters Summer flooding of the Nile




I 6-7 I4 I5 III 6 III 2-3

Comets Shooting stars Other luminary phenomena Rods and mock suns Halo


II 2, 355b20-32 I 14 II 1-3

I 13 (II 3, 359a18-b22)

(II 8, 367a1-11)

II 7-8

I 13 & II 4-6


Aristotle Meteorology I-III

Milky way






5b19-23 5b26-30 5b19 & 23-26

5b30 - 6a16


5a11-14 5a14-21 5a21-23 5a23-24 4a19-23 4a26-27 4a27-32 4a32 - b1 4b1-5 4a23-24 4a25-26 4a25 4b7 - 5a10


(5a32) & 5b8-9 (5a32) 5b3-17 5a35-36 5a36 - b3

t å met ãrsia


t å prÒsgeia

16 17

IV 1

(Stob. 39 ?)



7 8

4 4 4 4

3 3 3 3


1 2 2 2 6 18

Aëtius Placita III (+ IV 1)

[Aristotle] De Mundo 4 (394a19 - 6a27)

166.9-11 201-206a 222b 212-220 207

154-190 & 242-248 207 & 236-8 207-8 233-234a 224b-232



111 & 114-130

150b-151 112-3 112-3 112-3 & 135-46 131-4 152 111 & 152 111 152 152 152 152


89-94 96a 96b-101 99a 98a


Pliny NH II 89-248


III 4-5

(VI 4.1, etc.) VI 28 & III 21 III III 20 & 25-26 IVa





I 3-8

VII I 1 & 14-15 ? I 9-13 I2


Seneca NQ


2 1


6 9 8

11 10 12 13


4 5

t å §n é°ri gin.

Stoics ap. Diog. Laërt. VII 151-4



7 8 9 10 11 12

1 2-5 6 13


Syriac meteorology


16 & 36

17A 17B 26 25 27A 27B 28 24 (?)

18 19-20 21 22



31 35

Epicurus Pyth. 16-36

Table 2-1: Range of subjects and subdivisions in various ancient accounts of meteorology

906-1089 1090-1286


639-702 738-839 840-847 848-905 712-737


(527-534) (527-534) (527-534)

451-494 495-523 (527-534) (527-534)

96-159 160-203 204-422 423-450


Lucretius DRN VI 96-end




2.2.4 Some observations Fiery phenomena of the upper atmosphere In Meteor. I 4-8 Aristotle sets one group of phenomena apart as belonging to the fiery upper part of the atmosphere. To this group he assigns comets, shooting stars, and the Milky Way. These three phenomena are variously treated in subsequent works on meteorology. The Milky Way seems to have been excluded from this group quite early in the tradition.243 Except for Aëtius, all subsequent ‘meteorologists’ have omitted the subject. There seems to have been some doubt about the assignation of comets to this group as well, a doubt reported and shared by Seneca.244 Most meteorologists, however, were happy to follow Aristotle’s lead, discussing both shooting stars and comets under the general heading of atmospherical or lofty phenomena. The only explicit exception is Pliny, who classes comets and shooting stars together with some other luminary phenomena among astronomical matters. The positions of the Syriac meteorologist, Epicurus and Lucretius are harder to ascertain. Comets and shooting stars are not discussed in the Syriac meteorology and Lucretius book VI, which suggests that they fell outside the scope of these works. It must be noted, however, that both subjects are also absent from Lucretius’ astronomical passage in book V. Epicurus does not formally distinguish between astronomical and atmospherical phenomena, both of which he calls ‘lofty matters’ (met°vra). Yet, the structure of his Letter to Pythocles, at least up to chapter 29, indicates that he accepted at least a practical division between the two groups of phenomena. Unfortunately, the confused order at the end of the letter, where comets and shooting stars are discussed together with a number of unmistakably astronomical phenomena makes it hard to decide to which group Epicurus would have assigned comets and shooting stars, although the placement of the chapters on these subjects suggests that Epicurus associated both phenomena with astronomy (see p.87 above). Non-substantial luminary phenomena Another sub-class of atmospherical phenomena distinguished by Aristotle is that of the non-substantial luminary phenomena.245 The most notable of these are rainbows, ‘rods’, mock suns and haloes. Most subsequent meteorologists follow Aristotle and assign these to atmospherical or lofty 243

See note 197 above. See note 229 above. 245 On the ultimate Aristotelian origin of the distinction between substantial and nonsubstantial phenomena see Mansfeld (2005). 244


phenomena. Again, Pliny is the only explicit exception. He splits up the group, assigning mock suns (§99) and haloes (§98) to astronomy, and leaving only the rainbow (§§150-151) among atmospherical phenomena. In this case the Syriac meteorology seems to follow the majority view: among the otherwise meteorological phenomena it also includes the halo. It is strange, however, that the rainbow, which is the best known of this class of phenomena, should not have been included. In this respect the Syriac meteorology differs from all the other meteorologies. Lucretius’ account in book VI only mentions the rainbow, but omits the halo. Epicurus, in the Letter to Pythocles, discusses both the rainbow and the halo. The fact that he discusses them in consecutive chapters suggests that he too, like most other meteorologists, considered these two phenomena to be related and hence to belong to the same class of phenomena, i.e. atmospherical phenomena. Problems pertaining to the earth as a whole Aëtius’ section on tå prÒsgeia contains a number of chapters relating to the earth as a whole: (9) on the earth {being unique and limited}, (10) on the earth’s shape, (11) on the earth’s position, (12) on the earth’s inclination, (13) on the earth’s motion {or immobility}, and (14) on the earth’s division {into five zones}.246 The only other work to deal with such subjects within the scope of meteorological phenomena, is Pliny’s Naturalis Historia. These subjects are absent from all the other works in the table and we have explicit information that most of their authors considered such subjects cosmological and astronomical rather than meteorological: Aristotle discusses the shape, position and stability of the earth in his cosmological and astronomical work De caelo (II 13-14), and the same subjects, as well as the earth’s size and division into zones, are dealt with in the cosmological and astronomical treatise of the Stoic Cleomedes (I 1 & 5-8); Lucretius discusses the stability and location of the earth in the astronomical section of DRN book V (534563), and Epicurus dealt with the same subject in book XI of his magnum opus On nature (fr.42 Arr.), amidst a number of cosmological and astronomical problems.247 Seneca too, in the introduction to book II of the Naturales Quaestiones, tells us that certain problems concerning the earth, like its position, belong not to the terrena or sublimia but to the caelestia.248 It would appear therefore that the inclusion of such problems among otherwise meteorological phenomena is an innovation by Aëtius and Pliny, probably inspired by their wish for a rigorous top-down presentation of natural 246

The words between {…} are not part of the Aëtian chapter titles, but have been added by myself to better specify the precise subject of each chapter. 247 See Sedley (1998a) 119-21. 248 See n.231 above.



phenomena.249 In general, then, such problems do not belong to meteorology.250 Earthquakes In the table, earthquakes are variously placed among terrestrial or atmospherical phenomena. In the De mundo 4, in Aëtius III and in Pliny II, earthquakes are dealt with under the general heading of terrestrial phenomena. The reason for this seems to be that in all three works precedence is given to the location of the phenomenon. Seneca and the Stoics, on the other hand, agree in classing earthquakes with lofty or atmospherical phenomena. Seneca provides us with the reason for this – perhaps – surprising move (NQ II 1, 3): “Quomodo,” inquis, “de terrarum motu quaestionem eo posuisti loco quo de tonitribus fulguribusque dicturus es?” Quia, cum motus spiritu fiat, spiritus autem aer sit agitatus, etiamsi subit terras, non ibi spectandus est; cogitetur in ea sede in qua illum natura disposuit.

“Why,” you ask, “have you put the study of earthquakes in the section where you will talk about thunder and lightning?” Because, since an earthquake is caused by a blast, and a blast is air in motion, therefore, even if air goes down into the earth, it is not to be studied there; let it be considered in the region where nature has placed it’.251

According to Seneca, who may be supposed here to speak on behalf of all Stoics, earthquakes, being caused by air, should be dealt with in connection with other phenomena of the air. In this respect the Stoics follow closely in Aristotle’s footsteps. Although Aristotle does not yet apply the neat bipartion of meteorological phenomena into those of the earth and those of the air (or ‘lofty’: metãrsia / sublimia), such as we find with many of his successors, he does explicitly link the subject of earthquakes with that of winds (Mete. II 7, 365a14-15): Per‹ d¢ seismoË ka‹ kinÆsevw g∞w metå taËta lekt°on: ≤ går afit¤a toË pãyouw §xom°nh toÊtou toË g°nouw §st¤n.


After the previous subject (i.e. wind) we must speak about earthquakes and earth tremors: for the cause of this phenomenon is akin to that of wind.

On this top-down presentation of cosmological problems in Aëtius and other writers, and its consequences for the location of the sections dealing with the earth as a whole, see Mansfeld & Runia (2009a) 40-1 with n.71, and 133-4. On the order of Pliny’s cosmology see Kroll (1930) p.2, and Hübner (2002). 250 Similar observations and conclusions in Mansfeld (1992b) n.124 and Mansfeld & Runia (2009a) 119-22. 251 Transl. Corcoran (1971/2), slightly modified.


For the Syriac meteorology, and the accounts of Epicurus and Lucretius the story is a bit different. In all three works earthquakes are accounted for by a number of alternative explanations, not just wind, which makes their link to atmospherical phenomena less obvious. Yet, it can hardly be a coincidence that just as in the account of Stoic meteorology, so too in the Syriac meteorology and in the Letter to Pythocles, earthquakes are the only ‘terrestrial’ phenomenon to be discussed among a number of otherwise atmospherical phenomena. It would appear that even though the Syriac meteorologist and Epicurus do not share the Stoics’ and Aristotle’s assumptions, they do follow the tradition that incorporates earthquakes among atmospherical phenomena. It seems reasonable to suppose that Lucretius, being a follower of Epicurus, whose meteorological account, moreover, closely matches the Syriac meteorology, belongs to this same tradition. Terrestrial phenomena (other than earthquakes) In most of the meteorological accounts a number of terrestrial phenomena (other than earthquakes) are included. Lucretius, too, discusses a number of terrestrial phenomena. It is remarkable that the two closest parallels to Lucretius, viz. the Syriac meteorology and Epicurus’ Letter to Pythocles, do not deal with this class of phenomena. It could be and has been argued, in view of the close similarity between Lucretius’ book VI and the Syriac meteorology, that the latter must originally have dealt with such subjects too.252 There is no reason to assume that the Letter to Pythocles is incomplete, but Epicurus might have dealt with terrestrial phenomena somewhere else, perhaps in his On nature. Yet, there is no direct evidence for this claim, and the only terrestrial phenomena that we know for certain to have been discussed by Epicurus are magnets and diseases.253 It might therefore be claimed just as well that the section on ‘terrestrial phenomena’ in DRN VI was Lucretius’ own invention. In §2.3 below I will investigate how Lucretius’ account of terrestrial phenomena in the second half of book VI relates to his account of atmospherical phenomena in the first half, and also how it relates to the discussion of similar matters in other meteorological works and summaries.

2.2.5 Some conclusions The Syriac meteorology The range of phenomena covered in the Syriac meteorology is smaller than in any of the other meteorologies. For this reason it has often been argued that the treatise must be an extract from a larger work in which more subjects were 252 253

Steinmetz (1964) 216 with n.3; Mansfeld (1992a) 315-7. See p.98 with n.258 below.



dealt with. Steinmetz, for instance, suggests that the original work would have included chapters on the rainbow, mock suns, dew and volcanoes.254 Many of the arguments for its completeness or incompleteness are based on its identification with Theophrastus’ lost 2-volume Metarsiologika. Mansfeld, for instance, following a suggestion made by Daiber, believes that the chapter on earthquakes may originally have been the first chapter of Theophrastus’ second book, which, in addition to this chapter, ‘may have included treatment of other so-called terrestrial phenomena, e.g. “the advances and regressions of the sea and the extensions of the land” ’.255 Now let us for a moment forget the ascription to Theophrastus and compare the work as we have it with other meteorologies. As we saw above (p.94), the Syriac meteorology, just like D.L.’s account of Stoic meteorology and Epicurus’ Letter to Pythocles, includes earthquakes but excludes other terrestrial phenomena. Now, the reason for this practice in Stoic meteorology is clear: its subject matter is confined to ‘things happening in the air’ (tå §n é°ri ginÒmena); earthquakes are caused by moving air, and therefore must be classed with phenomena in the air. This motive, however, is not valid for Epicurus, for whom subterranean winds are only one of several explanations. The inclusion of earthquakes (Pyth. 23 [105-106]) among Epicurus’ met°vra (lofty things) can therefore only be explained by his willy-nilly adherence to a tradition, whose principles he no longer ascribes to. The same explanation can be applied to the Syriac meteorology. The inclusion of earthquakes (ch.15) in no way entails the inclusion of other terrestrial phenomena. Another group of subjects that appears to be missing is that to which comets and shooting stars belong. From Aristotle onwards these two subjects appear to have been standard ingredients of meteorological treatises. We find them included under atmospherical / lofty phenomena in the De mundo 4, in Aëtius book III, and in the account of Stoic meteorology in Diogenes Laërtius. As we have seen above,256 Seneca prefers to see comets as astronomical rather than meteorological phenomena, but in doing so he also testifies that comets were a traditional part of meteorology. Pliny is the only author who explicitly classes both comets and shooting stars under the general heading of astronomy. We do not know why the Syriac meteorologist omits both subjects, but he is not alone in doing so: they are also missing in Lucretius’ meteorological survey in book VI of the DRN. It is possible that both authors, like Pliny, considered these subjects astronomical rather than meteorological, 254

Steinmetz (1964) 216 with n.3. Mansfeld (1992a) 315-7. 256 See note 229 above and text thereto. 255


but this cannot be proved, as comets and shooting stars are equally absent from the astronomical section in DRN V. Both subjects are discussed in Epicurus’ Letter to Pythocles. However, the fact that they are discussed there in close association with a group of undeniably astronomical subjects may suggest that Epicurus thought them astronomical rather than meteorological (see p.87 above). However this may be, the absence of comets and shooting stars from a meteorological treatise is not unparallelled, and there is no need to suppose that the Greek original of the Syriac meteorology did discuss these subjects. This leaves us with only one more omission: the rainbow. The Syriac meteorology is the only meteorology which does not include the rainbow.257 This is the more striking as the treatise does discuss the halo, with which the rainbow is traditionally associated. If, for the sake of brevity, one of the two is omitted it is usually the halo: this is the case in Diogenes’ account of Stoic meteorology, and Lucretius book VI. In the meteorological section of Epicurus’ Letter to Pythocles, whose range of subjects is otherwise very close to that of the Syriac meteorology, both the rainbow and the halo are discussed. There seems to be some reason, then, to suppose that the Syriac meteorology, or its Greek source, may have contained a chapter on the rainbow, which was lost in the course of the transmission of the text. It appears to me, therefore, that, as far the range of subjects is concerned, except perhaps for its omission of the rainbow, the Syriac meteorology may well be complete. What does all this mean for the treatise’s attribution to Theophrastus? Very little, I am afraid. As the table shows, in its range of subjects it most closely resembles Diogenes Laërtius’ Stoics, the latter part of Epicurus Letter to Pythocles and the first half of the sixth book of Lucretius’ DRN. The omission of comets and shooting stars is parallelled in Lucretius VI, and perhaps also Epicurus’ Letter, if comets and shooting stars are taken to be astronomical (see p.87 above). The most telling argument in favour of a Peripatetic origin is its inclusion of earthquakes within a range of otherwise atmospherical phenomena. This fact suggests a dependence on a tradition in which the explanation of earthquakes was closely linked to that of wind. This tradition is most clearly exemplified by Seneca and the Stoics, and can be traced back to Aristotle. In this respect, however, the Syriac meteorology does not differ from Epicurus’ Letter to Pythocles; both works stand in the same relationship to the (Aristotelian) tradition, and there is no reason why one should be closer to the origin of the tradition than the other.


See Mansfeld (1992a) 315-16.



Epicurus’ Letter to Pythocles Most that can be said about the Letter has been said already: it does not explicitly differentiate between astronomical and meteorological phenomena, yet in the organization of the Letter some kind of a division appears to be present. At the end of the Letter the order of subjects is a bit confused, and due to this confusion it is not entirely clear whether Epicurus considered comets and shooting stars astronomical or rather atmospherical. The evidence seems slightly in favour of the first option (see p.87 above), and if this is true the range of truly atmospherical phenomena in the Letter would correspond almost exactly to that of the Syriac meteorology (except for the rainbow which is absent from the Syriac as we have it). Lucretius’ DRN book VI Lucretius’ meteorological account differs from Epicurus’ in two important respects. Firstly, whereas Epicurus does his best to obscure the difference between astronomical and meteorological phenomena by discussing them under the single heading of met°vra, Lucretius deals with both fields separately. The astronomical passage in book V 509-770 is firmly separated from the discussion of meteorological phenomena in book VI by the intervention of a very long section on the origins of life and civilisation, which occupies the second half of book V (771-1457). Secondly, the range of meteorological subjects discussed by Lucretius is considerably longer than that of either Epicurus’ Letter to Pythocles, or the Syriac meteorology, which in many other respects appears the closest parallel to Lucretius’ book VI. Whereas the Letter to Pythocles and the Syriac meteorology confine themselves to atmospherical phenomena and earthquakes, Lucretius proceeds to deal (608-1286) with a range of (other) terrestrial phenomena, such as the sea, the Etna, the Nile, poisonous exhalations, springs and wells, magnets and diseases, which are not represented in the Letter to Pythocles or in the Syriac meteorology. It cannot be excluded, of course, that Epicurus dealt with such matters elsewhere: Galen credits him with an elaborate theory of magnetism,258 and Diogenes Laërtius ascribes to him a work titled Per‹ nÒsvn dÒjai prÚw M¤yrhn, Opinions on diseases, to Mithres,259 which may or may not have dealt with the physical side of diseases. However, there is no evidence that he discussed any of the other subjects which Lucretius covers in DRN VI 608ff. It is possible, therefore, that Epicurus was not Lucretius’ main source for this 258 259

Galen, On Natural Faculties, I 14 [vol. II p. 45 Kühn] (= Epic. fr.293 Us.). Diog. Laërtius X 28.


passage. In the next section the character of Lucretius’ passage and its possible relations to meteorological and other literature will be examined.

2.3 Terrestrial phenomena other than earthquakes 2.3.1 Lucretius On p.89 above it was suggested that Lucretius’ section on the constant size of the sea was the first of a new class of problems, different in character from the preceding phenomena, both atmospherical and earthquakes. The clue as to what this difference might be is given right at the beginning of this new division (lines 608/9): ‘Principio mare mirantur non reddere maius / naturam, ...’ – ‘In the first place people wonder why nature doesn’t make the sea bigger, ..’ As Giussani and Bailey point out, this sense of wonder also characterises many of the problems that follow (explicitly so in 608 mirantur, 654-55 mirari & miratur, 850 admirantur, 910 mirantur, 1056 mirari). In this sense, then, these problems differ from the preceding ones, which may incite awe and fear, but are not said to cause wonder.260 Although Lucretius does not tell us explicitly in what way the second group of phenomena should inspire this sense of wonder, which the preceding do not, it is not hard to see that there is a difference in character between the phenomena in the first group and most of the second group. While the phenomena in the first group (including earthquakes) are all capable of occurring just about anywhere, the majority of the subjects discussed in the second part are concerned with exceptional and local phenomena, the kind of phenomena the ancients referred to as parãdoja, yaumãsia or yaÊmata and mirabilia or miracula,261 i.e. ‘marvels’ or ‘miracles’.262 Such are the Etna, the river Nile (explicitly said to be ‘unique’ - 713 unicus), the ‘Avernian’ places (one near Cumae, one in Athens and one in Syria), and the spring near the shrine of Hammon (in the Siwa-oasis in Egypt). Also local is the cold spring which kindles tow (lines 879-905), whose location Lucretius does not reveal, but which may be identified with either the spring of Jupiter in Dodona or the


This is not entirely true: in the introduction to book VI Lucretius speaks of people wondering (59 mirantur) about things that take place above our heads in the ‘ethereal’ regions (61 quae supera caput aetheriis cernuntur in oris), i.e. astronomical and atmospherical phenomena. Yet, in the body of the text the use of this verb is restricted to certain terrestrial phenomena only. 261 For ancient names for such phenomena see Ziegler (1949) cols. 1137-38; Schepens & Delcroix (1996) 380-2; Wenskus (2000) col. 309. 262 Lists of such phenomena are found throughout Pliny’s Naturalis Historia, where they are referred to as miracula and mirabilia.



spring of the nymphs in Athamania,263 about which similar stories were told. The magnet, too, may be counted among local phenomena, as it is found specifically – so Lucretius tells us – in the land of the Magnetes (in Lydia, Asia Minor). I have summarized all this in the following table (for the sake of completeness I have also included 703-711 which do not deal with a specific phenomenon, but with the method of multiple explanations in general): Table 2-2: Lucretius’ account of terrestrial phenomena lines 608-638 639-702 703-711 712-737

840-847 848-878 879-905

subject Why the sea does not grow bigger Etna Multiple explanations The summer flooding of the Nile Lacus Avernus (745-47) Avernian Acropolis (748-55) places Syria (756-59) Water in wells colder in summer The spring at the shrine of Hammon A cold spring which kindles tow


The magnet


1090-1286 Diseases

exceptional 608 mirantur 654 mirari


713 unicus

712 Nilus 747 Cumas apud 749 Athenaeis in moenibus 756 in Syria

850 admirantur

848 apud Hammonis fanum (Dodona / Athamania)

910 mirantur 1056 mirari

639 Aetna

909 Magnetum in finibus 1115 Aegypto 1116 Achaeis finibus 1117 Atthide

Three phenomena stand out in the above list: the constant size of the sea, wells being colder in summer, and diseases. The sea, which occupies such a large portion of the earth,264 can hardly be called a local phenomenon. It is clear that the sense of wonder it it said to inspire is of a different kind from that inspired by, for instance, the Nile or the spring of Hammon. Thematically it seems to be more closely related to the atmospherical phenomena of the preceding section: one of the explanations offered (627-630) – viz. that a considerable portion of water is drawn up by the clouds – is the exact counterpart (as Lucretius himself points out in 627) of one of the causes of cloud formation (470-475) and of rain production (503-505).265


See Ernout-Robin (1925-28) ad loc. The spring of Jupiter in Dodona is described by Pliny NH 2.228.1-3 and Mela 2.43 and the spring of the nymphs in Athamania by Antigonus 148, the Doxographus Florentinus 11 and Ovid Met. 15.311-12. 264 Cf. Lucr. DRN V 203. 265 For this reason Robin (Ernout-Robin ad loc.), ignoring Lucretius’ own textual clues, prefers to include this passage with the atmospherical phenomena that precede it.


The account of why water in wells is colder in summer than in winter is also different. Whereas most phenomena in this section are somehow exceptional among their kind – the Etna among mountains, the Nile among rivers, the spring of Hammon among springs and the magnet among stones –, this passage is about something generally attributed to all of its kind: all wells were believed to be colder in summer and warmer in winter.266 Yet it is not hard to imagine why Lucretius included it in this section: the annual temperature fluctuation of wells is somewhat similar to the daily temperature fluctuation of the spring of Hammon. This brings us to the last subject, not just of the ‘terrestrial phenomena’ but of the entire book: diseases. The language with which Lucretius introduces the subject does not suggest any major break with the preceding subjects. Soon, however, it appears that they are somewhat different. Diseases, according to Lucretius, are produced in two different ways (1098-1102): Atque ea vis omnis morborum pestilitasque aut extrinsecus, ut nubes nebulaeque, superne per caelum veniunt, aut ipsa saepe coorta de terra surgunt, ubi putorem umida nactast intempestivis pluviisque et solibus icta.

And all this might of diseases and this pestilence either comes from without, like clouds and mists, from above through the sky, or often, having gathered, they rise from the earth itself, when this, being moist, has come to rot, having been hit by out-of-season rains and suns.

Diseases either come from without through the sky,267 like clouds and mist, or they arise locally from the earth itself. These two kinds seem to correspond to what ancient as well as modern medicine refers to as epidemic and endemic diseases.268 In the following lines (1103-1118) Lucretius first deals with the second kind of diseases: those that are peculiar to certain regions and peoples, depending on the local climate, and which may also affect those who travel 266

Cic. N.D. II 25.7 - 26.1; Sen. NQ VI 13.3-4; Plin. NH II 233.1-2. This theory is anticipated in line 956 (‘morbida visque simul, cum extrinsecus insinuatur’) in the account of the magnet, and seems to look back to 483ff where the possibility of an extra-cosmic origin of clouds is suggested. Line 956, which explicitly refers to disease, seems to ascribe an extra-cosmic origin to diseases as well, but only if it is connected with the preceding line from which it is separated by a lacuna of unknown length (see the commentaries). In the present passage, however, apart from the reference to clouds, there is nothing to suggest that diseases might come from without the cosmos. The point rather seems to be that diseases are either innate to a certain region, or come from elsewhere. See Kany-Turpin (1997). 268 Bailey does observe the distinction but makes nothing of it. He calls all the diseases in this passage ‘epidemic’ and also ‘marvel’. For the distinction between epidemic and endemic diseases, see Galen In Hippocratis de victu acutorum commentaria iv, Kühn 15.429-430; id. In Hippocratis librum primum epidemiarum commentarii iii, Kühn 17a.13; Ps.-Galen Definitiones medicae, Kühn 19.391. See also the admirably concise accounts in Karl-Heinz Leven (ed.) Antike Medizin. Ein Lexikon, Munich 2005, lemmata ‘Endemie’ and ‘Epidemie’. 267



there. Lucretius ends this brief account with three examples that would not be out of place in our list of local marvels: elephantiasis which is unique to Egypt, a foot-disease peculiar to Attica and an eye-disease typical of Achaea. Then (1119-1132) Lucretius goes on to discuss the other kind of diseases: those that travel with the air like clouds and mist and thus come upon us. Before coming to the ‘finale’, Lucretius recapitulates the major point of his account (1133-1137): it makes no difference whether we travel to an unwholesome place (endemic disease) or whether nature brings the unwholesomeness to us (epidemic disease). Lucretius ends the account, and the entire work, with a long description of an outbreak of the second type (haec ratio morborum), the famous plague of Athens.269 In sum, Lucretius’ account of diseases does not entirely blend in with the preceding accounts of mostly local marvels. While endemic diseases bear a certain resemblance to local marvels such as the Nile, the spring of Hammon and especially Avernian places, which are accounted for in a similar manner,270 epidemic diseases have more in common with the non-local atmospherical phenomena of the first half of book VI; they are repeatedly likened to clouds and mist. Beside these three problems – the constant size of the sea, the paradoxical annual temperature fluctuation in wells, and diseases – all the other subjects discussed in the second half of book VI are concerned with exceptional local phenomena.

2.3.2 Local marvels in meteorology and paradoxography Although some of the local marvels discussed by Lucretius, as well as similar ones, are also mentioned in various other meteorologies, phenomena of this kind are more typically found in an another genre of writings, generally known as paradoxography.271 Paradoxography is the activity and the written result of collecting accounts of natural marvels. Such ‘marvels’ comprise ‘unexpected features of the natural world (animals, plants, rivers and springs), but also the world of man, human physiology, unusual social customs, and even curious historical facts (...)’,272 which are drawn from all kinds of earlier writings, often with explicit 269

In ancient medical literature (e.g. Galen In Hippocratis de victu acutorum commentaria iv, Kühn 15.429) a plague (Greek loimÒw) is defined as a deadly epidemic (§pidhm¤a Ùl°yriow). See also Karl-Heinz Leven, op. cit., lemma ‘Pest’. 270 See esp. lines 769-780 on the earth containing both beneficial and harmful elements; and lines 781-817 listing a number of harmful substances and places. 271 On paradoxography in general see e.g. Ziegler (1949); Schepens & Delcroix (1996). 272 Schepens & Delcroix (1996) 381.


acknowledgement of the source.273 Paradoxographies do not usually express doubt about the veracity of the marvellous stories they report, nor do they provide physical explanations for them; they restrict themselves to simply reporting them. In this respect paradoxography differs from the scientific writings from which many of the marvellous stories were culled. Discarding the On marvellous things heard, which is generally believed to be the work, not of Aristotle, but of a number of subsequent authors working between the 3rd cent. BC and 2nd cent. AD,274 the oldest reported writer of paradoxography is Callimachus of Cyrene (ca. 310-240 BC), part of whose work is reproduced in Antigonus’ Historiarum mirabilium collectio.275 The place of natural marvels in ancient scientific literature – and especially meteorology – and the relationship between science and paradoxography are complex subjects. Although paradoxography as a separate genre seems to have originated only in the Hellenistic age, scientific works like Aristotle’s not only provided it with many individual marvel stories, but actually set the example of producing lists of particular local phenomena. Aristotle’s Meteorology, for instance, contains two lists of particular waters: at 350b36 351a18 of partially underground rivers, and at 359a18-b22 of salty and othertasting waters, many particular instances of which are also reported in paradoxographical literature. In its turn scientific literature of later ages reappropriated a lot from paradoxography. In book II of Pliny’s Naturalis historia, for instance, (as in the rest of his work) several long lists of ‘miracula’ (so Pliny calls them) are produced, which are almost indistinguishable in character from the accounts found in purely paradoxographical works, from which they seem to have been borrowed.276 Both the attitude towards, and the space devoted to, particular local phenomena in ancient meteorologies changes in the course of time. Aristotle’s attitude towards such phenomena is ambiguous. Although, as we have just seen, the Meteorology includes several lists of such particular problems, these are never the actual objects of inquiry. The list of underground rivers at 350b36 - 351a18 only serves to illustrate why some people might – incorrectly – think that the sea itself is replenished from underground reservoirs, and the list of salty waters at 359a18-b22 serves to lend plausibility to the belief that the salty taste of sea-water too is due to admixture. In later meteorologies the amount of paradoxographical passages increases dramatically. In the De mundo the entire section on terrestrial phenomena, save for a long account of 273

Schepens & Delcroix (1996) 383-6. Ziegler (1949) cols. 1150-51; Schepens & Delcroix (1996) 427; Vanotti (2007) 46-53. 275 Antigonus Hist. mir. 129ff. 276 On Pliny’s paradoxographical passages see e.g. Ziegler (1949) cols. 1165-66; Schepens & Delcroix (1996) 433-9; Naas (2002), ch.5. 274



earthquakes, is a mere enumeration of local marvels. In Seneca’s Naturales Quaestiones a large portion of book III ‘On waters’ is filled with lists of peculiar waters. Pliny’s Naturalis Historia caps them all: book II contains extensive lists of earth miracles (206b-211), water miracles (224b-234) and fire miracles (235-238), often, as in paradoxography proper, with explicit reference to the source of each story. As a rule, in paradoxography as well as meteorology, such peculiar local phenomena are not explained, or only in the most general terms. A curious exception to this rule is the summer flooding of the Nile. True to his own precept, in Metaph. VI 3, 1027a20-26,277 that science should not be concerned with particular, but only general problems,278 Aristotle does not even mention the Nile flood in his Meteorology, where we only find the general and unqualified observation (I 13, 349b8) that all rivers flow higher in winter than in summer.279 Interestingly, among the works ascribed to Aristotle there is a treatise, only preserved in a 13th century Latin translation, titled Liber de inundacione Nili (‘Book on the flooding of the Nile’), which is entirely devoted to the question of why the Nile, in contrast to all other rivers, overflows in summer.280 Although the work’s authorship is still a matter of dispute, most scholars agree that it should at least be assigned to Aristotle’s school.281 Apparently then, even in Aristotle’s school such a particular problem could be worthy of scientific investigation after all. It need not surprise us, therefore, that in some subsequent meteorological accounts the problem of the Nile flood was simply included within the scope of their subject matter. This is the case with Aëtius and Seneca, who devote a whole chapter (IV 1) and a whole book (IVa) to it, respectively. It is interesting to note that precisely this problem, perhaps the most famous of all local marvels, 277

Cf. An. post. I 8, 75b21-36. See Taub (2003) 83. 279 Arist. Mete. I 13,349b8: DiÚ ka‹ me¤zouw {sc. toÁw potamoÁw} toË xeim«now re›n µ toË y°rouw ... – ‘Therefore also (they suppose that) rivers flow higher in winter than in summer …’. 280 Arist. (?) Liber de Nilo 2-4: ‘Propter quid aliis fluminibus in hyeme quidem augmentatis, in estate autem multo factis minoribus, {sc. Nilus} solus eorum, qui in mare fluunt, multum estate excedit ... ?’ – ‘For what reason, while other rivers rise in winter, and become much smaller in summer, does the Nile, alone of those that flow into sea, rise strongly in summer ... ?’ 281 The work is variously attributed to Aristotle and Theophrastus. An influential case for Aristotle’s authorship has been made by J. Partsch (1909); and for Theophrastus’ by P. Steinmetz (1964) 278-96. Both views have had their adherents until quite recently; for an overview see Sharples (1998) 197 with notes. Partsch’s view is also accepted by R. Jakobi & W. Luppe (2000). 278


does not feature in any of the surviving paradoxographies. Perhaps the very fact that so many had studied it, and provided explanations for it, disqualified it as a ‘mirabile’. In what follows I would like to view Lucretius’ account of – mostly local and particular – terrestrial phenomena against the background of the treatment of such phenomena in other meteorological as well as paradoxographical works. As some of the ‘meteorologies’ we have hitherto referred to do not discuss terrestrial phenomena (except for earthquakes) at all, I will here limit myself to those that do, viz. Aristotle’s Meteorology, Pseudo-Aristotle’s De mundo 4, Aëtius’ Placita III, Seneca’s Naturales Quaestiones and Pliny’s Naturalis historia II. From among the many paradoxographical works dealing with terrestrial phenomena, I have chosen to include in the table only Antigonus’ Historiarum mirabilium collectio and Pseudo-Aristotle’s De mirabilibus auscultationes, which are the most extensive ones and present most parallels. In the accompanying text occasional references shall be made to other paradoxographies as well, such as Apollonius’ Historiae Mirabiles, Claudius Aelianus’ De natura animalium and the so-called ‘Paradoxographus Vaticanus’ and ‘Paradoxographus Florentinus’. In the table below I have included all terrestrial phenomena discussed by Lucretius in DRN VI 608ff, to which I have added the sweet spring in the sea off Aradus, which Lucretius mentions (890-894) as a partial parallel to the spring that kindles tow (879905), and which may well derive from the same stock. I have also added two more general categories, volcanoes and marvellous waters, of which Lucretius offers only some specimina. In the table a reference in bold characters indicates that the phenomenon in question is not merely mentioned but also physically accounted for. Brackets indicate that the phenomenon is not mentioned in its own right but only to serve as a parallel to, or an example of, some other phenomenon. A wavy line (~) indicates an almost identical phenomenon at a different geographical location. References to the De mundo 4 are indicated by the last digit of the Bekker page only, the first two digits of the relevant pages being always 39.


CHAPTER TWO Table 2-3: The subjects of DRN VI 608ff with parallels in meteorology and paradoxography. Work

Subject Constant size of the sea Volcanoes - Etna The Nile flood

Lucretius DRN VI 608-638 639-702 712-737

Poisonous exhalations


a. Lacus Avernus b. Acropolis c. Syria / Phrygia (*)

746-748 749-755 756-759

Temperature in wells


Aët. Plac. III

[Arist.] De mundo 4 5b19-23 5b21

IV 1 5b26-30

848-878 879-905 (890-894) 906-1089 1090-1286

Seneca NQ III 4-8 (V 14.4 etc.) IV a (III 21) & (VI 28)


(350b36 ff) (359a18 ff)

Marvellous waters - Spring of Hammon - Spring which kindles tow - Sweet spring off Aradus Magnets Diseases

Arist. Meteor. I-III 355b20-32 (367a1-11)

Pliny NH II 166.9-11 236-238 236.1-3 207.9208.10



[Arist.] Mir.Ausc.


34-40 38b & 40

121-123, 152a-b 152b 12 123


53-57 etc.

(VI 13.3-4) & (IV a 26-27)


III 20 & 25-26



228.6-10 228.1-6 227.4-5

144 148 ~129.2

(~351a14-16) (VI 27-28)

(* On the identity of these two places see text on p.107 below)

Below I will briefly discuss each subject (or group of subjects), point out the parallels in meteorological and paradoxographical literature, and indicate to what extent Lucretius’ treatment corresponds to, and differs from, these. Constant size of the sea This problem is not, as we have seen, a ‘mirabile’ in the technical sense, and accordingly not found in any ancient paradoxography. Thematically the problem is closely related to atmospherical phenomena and was probably discussed in this context long before Aristotle’s Meteorology: in Aristophanes’ Clouds (produced ca. 420 BC), in a parody of contemporary physical theory,282 the constant size of the sea is one of several problems to be discussed (1278-1295), the others being the origin of rain (369-371), the causes of thunder (375-394), and the causes of the thunderbolt (395-407). Later the same problem is discussed in a meteorological context in Aristotle’s Meteorology and Seneca’s Naturales Quaestiones, and briefly touched upon in Pliny’s Naturalis Historia II. Lucretius’ discussion of this subject is entirely in line with this.


For the date see Dover (1968), pp.lxxx-xcviii. For the parody of contemporary physics see Dover’s and other commentaries ad locc.


Volcanoes Although earthquakes appear to have been a standard ingredient in ancient meteorologies, volcanoes were not.283 Aristotle only briefly refers to them, as possible side-effects of earthquakes, not phenomena to be studied in their own right, and he does not even mention the most formidable of them all, the Etna. Seneca’s attitude is similar. After some brief references to volcanic phenomena in book II (at 10.4; 26.4-6 and 30.1), in book V (14.4) he promisses to discuss the subject more fully in connection with earthquakes. Book VI on earthquakes, however, contains only one, disappointingly brief, reference to volcanoes, in a catalogue of possible side-effects of earthquakes (VI 4.1). Aëtius omits the subject alltogether. Things are different in the De mundo and in Pliny’s Naturalis Historia. In both works volcanoes are clearly set appart from earthquakes, and both works mention a number of volcanoes by name, including the Etna. Neither work, however, makes any attempt at explaining volcanism. In this respect they resemble the paradoxographical accounts in Antigonus (166-167) and the Pseudo-Aristotelian De mirabilibus auscultationes (34-40), only the latter of which includes the Etna itself. Lucretius’ approach is entirely different: he not only mentions the Etna as a phenomenon to be studied in its own right, but he actually explains its working. The Nile flood From very early on the summer flooding of the Nile had aroused the interest of the Greeks. Whereas, in their experience, all other rivers rose in winter, the Nile alone did so in summer. This called for an explanation and many different theories were devised to account for this curious behaviour.284 Aristotle does not deal with the subject in his Meteorology, but it is discussed at length in Seneca’s Naturales Quaestiones and in the meteorological section of Aëtius’ Placita. It would seem therefore that Lucretius, by including this local phenomenon, is perfectly in line with the meteorological tradition. Poisonous exhalations Lucretius next discusses what he calls ‘Avernian places’ (738 loca Averna), places where poisonous exhalations rise from the ground.285 Many 283

Hine (2002) 58-60. Different theories concerning the summer flooding of the Nile are listed by Herodotus Hist. II 19-27, [Arist.] Liber de Nilo, Seneca NQ IVa, Aëtius IV 1, and several others. For a conspectus see Diels (1879), 228. 285 Others referred to such places as ‘Charonia’ (Plin. NH 2.208; Strabo 12.8.17; Antigonus 123; Stoics SVF III 642 apud D.L. VII 123; etc.) or ‘Plutonia’ (Cic. Div. 1.79; Strabo Geogr. 5.4.5; etc.): cf. Lucr. VI 762 ‘ianua Orci’. There are no ancient parallels for Lucretius’ generic use of the word ‘Avernus’, unless perhaps, knowing that lake 284



instances of such phenomena are mentioned in paradoxography, and the subject also turns up in some of our ‘meteorologies’, such as the De mundo 4, Pliny’s Naturalis Historia II and Seneca’s Naturales Quaestiones. Seneca discusses the subject twice (at III 21 and VI 28) but gives no specific examples, and makes no real effort at explaining the phenomenon. The De mundo (395b26-30) offers a few examples, but no explanations. Pliny’s account (207.9-208.10) is by far the longest, and yet the least scientific. It consists of a long list of instances, which is itself part of an even longer list of phenomena which Pliny describes as ‘earth’s wonders’ (206.5 terrae miracula). Far from physically explaining these phenomena Pliny ends his account by simply attributing them to the ‘divine power of nature’ (208.9-10 numen naturae). Lucretius’ attitude is entirely different from Pliny’s: for Lucretius these phenomena, like any other, must and can be explained physically. Lucretius offers three examples. The first example is the Lacus Avernus, where overflying birds are said to fall dead from the sky. Although the case was well known in antiquity, and is reported in several paradoxographies (Antigonus 152b, Mir. ausc. 102, Paradox. Vat. 13), it is not mentioned in any of our ‘meteorologies’.286 Lucretius’ second example is the Athenian Acropolis which crows are said to avoid. This story, too, is not mentioned in any of our ‘meteorologies’,287 but it is found in several paradoxographies: Antigonus 12, Apollonius 9, and Claudius Aelianus’ Natura animalium V 8. None of these, however, relate this phenomenon to poisonous exhalations. The only ancient text, beside Lucretius’, that makes the connection is Philostratus’ Life of Apollonius of Tyana (II 10), where the absence of crows on the Acropolis is compared to the absence of birds on mount Aornus on the fringes of India, and to similar phenomena in Lydia and Phrygia. Lucretius’ third example, a place in Syria where four-footed animals collapse and die, is not known from any other ancient text. Lucretius’ description is, however, very similar to the stories told about a place near Avernus was called Aornos (bird-less) in Greek, Lucretius felt he could render other instances of the word ‘Aornos’ as ‘Avernus’ as well. Other places of a similar nature bearing the name ‘Aornos’ or ‘Aornon’ were reported to exist near Thymbria in Caria (Strabo, 14.1.11) and in Thesprotis in northern Greece (Pausanias, 9.11.6). Philostratus (Life of Apollonius of Tyana II 10) attributes a similar character to Mount Aornos in India. 286 Pliny knows the story (NH XXXI xviii 21), but does not mention it in his list of places with poisonous exhalation in NH II. 287 Pliny knows the story (NH X xiv 30), but does not mention it in his list of places with poisonous exhalation in NH II.


Hierapolis in Phrygia,288 and Robin (Ernout-Robin ad VI 749ff) even actually quotes Lucretius with ‘Phrygia’ instead of ‘Syria’, without commenting on the change. It is possible that Lucretius’ story is simply a garbled version of the reports about the site in Phrygia, transferred somehow to Syria, perhaps through confusion between the Phrygian Hierapolis and its Syrian namesake.289 Temperature in wells Lucretius next discusses the temperature variation in wells, which were commonly believed to be warm in winter, and cold in summer.290 This problem is not a ‘mirabile’ in the technical sense of the word, as it is not confined to one or a few specific places, but common to all of its kind. Accordingly it does not normally feature in paradoxographical works, although in Pliny’s NH II it is mentioned (233.1-2) as part of a long section on ‘water miracles’ (224b-234), consisting for the most part of true ‘mirabilia’. The problem is not generally discussed in meteorology either. Beside Lucretius and Pliny, the only ‘meteorological author’ who mentions the phenomenon is Seneca, who refers to it twice: in book IVa in the context of Oenopides’ account of the Nile flood,291 and in book VI in the context of Strato’s explanation of earthquakes. Curiously Seneca’s attitude varies: in book VI he accepts the observation as well as Strato’s explanation of it, but in book IVa he rejects not just Oenopides’ theory, but the observation itself: wells and other underground recesses only seem warm in winter and cold in summer, because they are protected from external temperature fluctuations. It is possible, as Robin observes,292 that the present subject’s connection with the Nile flood (as testified by Seneca and others) is what persuaded Lucretius to include it in his account of terrestrial phenomena.


Pliny NH II 208.4-5; Apuleius De mundo 17.17ff; Paradoxogr. Vaticanus 36, etc. Esp. this last account is very close to Lucretius’: ÉEn ÑIerapÒlei tÒpow §st‹ Xar≈niow legÒmenow, §n ⁄ oud¢n z“on d∞ta ba›nei: p¤ptei går paraut¤ka. ‘In Hierapolis there is a so-called ‘Charonian’ place, in which no animal goes: for it falls immediately.’ Note that in this account ‘Hierapolis’ is not further specified, so that someone unfamiliar with the story might easily connect it to the wrong city. 289 The Syrian Hierapolis, a.k.a. Bambyce, is mentioned by Plut. Ant. 37 & Crass. 17; Strabo 16.1.27; Pliny NH V 81; Ael. De nat. an. 12.2; and Pseudo-Lucian De dea Syria (passim). 290 See n.266 above 291 See also Diodorus Siculus I 41. 292 Ernout-Robin ad DRN VI 840-847.



Marvellous waters Lists of marvellous or peculiar waters were a standard ingredient of ancient paradoxography. Such lists are found for instance in Antigonus (129-165: ‘borrowed’, as Antigonus himself claims, from Callimachus), in the De mirabilibus auscultationes (53-57 et passim), and in the Paradoxographus Florentinus. Similar lists, of various lengths and serving various purposes, are also found in some of our meteorologies. In Aristotle’s Meteorology two such lists occur: in I 13, 350b36-351a18, Aristotle mentions a number of rivers that flow underground before emerging again, and in II 3, 359a18-b22 a number of salty and other-tasting waters; yet neither group is presented as something to be studied in its own right: the first group serves merely as a concession to those who believe that all rivers are fed from one or several underground reservoirs, a point of view rejected by Aristotle, and the second group serves to illustrate Aristotle’s theory that the salty taste of sea water is due to the admixture of something. Longer lists are found in book III of Seneca’s Naturales Quaestiones 20 & 25-26. Formally the purpose of Seneca’s lists is to illustrate that tastes and other peculiar properties of local waters are acquired through contact with some other substance, but the second list ends in a mere enumeration of marvel stories. Finally, Pliny, who offers the longest list (cvi 224b-234), does not even pretend that these are anything other than marvels: in his own table of contents this section is described as ‘Mirabilia fontium et fluminum’ (Marvels of sources and rivers), and similar terms recur throughout the passage.293 Lucretius’ approach is very different. Instead of producing a long list of marvellous waters, he singles out just two instances for further discussion: the spring of Hammon, and a spring (not otherwise identified) which kindles tow. A third spring, the spring of sweet water in the sea off Aradus, is not discussed but only described to serve as a partial analogy for the spring which kindles tow. Yet, as we shall see, it is likely that all three water marvels stem from the same source text. The first marvellous water described by Lucretius is the spring of Hammon whose waters are cold in the day-time and hot at night. This story, first reported by Herodotus and repeated by many historians and other authors since, is also found in Pliny’s section on marvellous waters in NH II (228.610), but not in any of the other meteorologies. A variant account, according to which the spring’s waters are hot at midnight and midday, but cold at


Pliny NH II 224 ‘mirabilius’, ‘miraculis’, 227 ‘natura mira’, 232 ‘mira oracula’, 233 ‘permira naturae opera’.


daybreak and sunset, is only found in the paradoxographical works of Antigonus (144) and the Paradoxographus Florentinus (19). Lucretius next gives an account of a cold spring which kindles tow that is held above. He does not identify the spring but his account is very similar to the stories told of the spring of Jupiter at Dodona, reported in Pliny’s section on marvellous waters (228.1-6) and in Pomponius Mela 2.37, and the spring at the shrine of the nymphs in Athamania, reported by Antigonus (148) and the Paradoxographus Florentinus (11). In the course of this account Lucretius also mentions the spring of sweet water in the sea near the island of Aradus off the coast of Phoenicia, a spring which – among the ancients – is only mentioned by Strabo (16.2.13) and in Pliny’s section on marvellous waters (227.4-5). In Pliny’s account it is connected to a similar spring near the Chelidonian islands off the coast of Lycia in Asia Minor. Of these two springs Antigonus only mentions the latter (129.2). Other springs of this nature are reported to exist in the eastern part of the Black Sea in Aristotle’s list of underground waters (Mete. I, 13.351a1416). Magnets This is the first among the ‘terrestrial phenomena’ that we know was discussed by Epicurus too.294 Before him the subject had been discussed by Thales, Empedocles and Democritus.295 Unfortunately we do not know the context of these discussions. The subject of magnets is not found in any of our ‘meteorologies’. The only link that I can see is with the subject of stones in general, which is briefly touched upon in Aristotle’s Meteorology III 6, and even more briefly in book II of Pliny’s NH 207 (see Table 2-1 on p.91 above). Surprisingly, the subject does not occur in any of the surviving paradoxographies either, although it would not have been out of place there. The closest parallel in paradoxography is the strange claim in Apollonius Hist. Mir. 23 that magnets attract iron only during the day-time. Perhaps, as with the Nile-flood, the fact that so many had already investigated and explained the phenomenon banned it from inclusion in paradoxographies. Diseases This is another subject that we know to have been discussed by Epicurus himself.296 There is no evidence, however, that Epicurus had somehow linked the subject to meteorology, as Lucretius does. 294

See note 258 above. Thales A22 D-K (Arist. De an. A2, 405a19ff), Empedocles A89 D-K (Alex. Aphr. Quaest. II 23), Democritus A165 D-K (Alex. Aphr. Quaest. II 23). 296 See p.98 above. 295



Above (on p.101) we have seen that Lucretius distinguishes two kinds of diseases: epidemic diseases, which come upon a place and after a while abate again, and endemic diseases which are peculiar to a certain place and never wholly disappear. Lucretius gives us three examples of endemic diseases (1115-7): elephantiasis in Egypt, a particular foot-disease in Attica and an eyedisease in Achaea. Although these and similar cases would not seem out of place in paradoxography, in fact they do not feature in any surviving work of that genre. Nor are diseases generally discussed in meteorological works. The only meteorologist, beside Lucretius, who has something to say about diseases is Seneca. In NQ VI 27-28, remarking upon certain peculiar phenomena that accompanied the Campanian earthquake of 62 AD, Seneca also comments on the ensuing death of hundreds of sheep, which he attributes to a plague. It is said, he claims, that plagues often occur after great earthquakes. This he accounts for on the assumption that the earth contains many harmful and lethal seeds, which may be released by the force of the earthquake. That the earth should contain such pestilential seeds can be inferred also, writes Seneca, from the existence of certain places that emit poisonous exhalations all the time (VI 28.1). So, according to Seneca, plagues are related in their origin to poisonous places, and may occur as a symptom of earthquakes. One might be tempted to see the account in Seneca’s NQ as an example of the kind of treatment that might have induced Lucretius to include diseases in a discussion of mostly meteorological phenomena. Yet, it seems more likely that the relationship went the other way: Seneca’s description of diseases and poisionous exhalations appears to owe a lot to Lucretius, who may well have been Seneca’s source of inspiration.297 But if Lucretius’ inclusion of diseases with meteorological phenomena was inspired by neither meteorology nor paradoxography, perhaps the link must be looked for outside these genres. An interesting parallel for the inclusion of diseases among a number of mostly meteorological occurrences is provided by Cicero. In ND II 13-15 (= SVF I 528), Cicero relates the four causes which, according to Cleanthes, are responsible for forming the notion of gods in the minds of men. The third cause (ND II 14) is described as follows:


Piergiorgio Parroni (2002), for instance, refers to Lucr. VI 1093-6 and VI 740-6 as possible sources of inspiration for Sen. NQ VI 27-28.

EPICURUS & LUCRETIUS AND THE SCOPE AND STRUCTURE OF ANCIENT METEOROLOGY 113 Tertiam quae terreret animos fulminibus tempestatibus nimbis nivibus grandinibus vastitate pestilentia terrae motibus et saepe fremitibus lapideisque imbribus et guttis imbrium quasi cruentis, tum labibus aut repentinis terrarum hiatibus tum praeter naturam hominum pecudumque portentis, tum facibus visis caelestibus tum stellis is quas Graeci komÆtaw nostri cincinnatas vocant ..., tum sole geminato, quod ut e patre audivi Tuditano et Aquilio consulibus evenerat ..., quibus exterriti homines vim quandam esse caelestem et divinam suspicati sunt.

As a third cause (Cleanthes posited that) which might frighten the minds on account of thunderbolts, storms, clouds, snow, hail, devastation, pestilence, earthquakes and frequent rumblings (of the earth), showers of stones, blood-like drops of rain, then landslides or sudden chasms in the earth, then preternatural portents of man and beast, then the sight of celestial torches, then those stars which the Greeks call kom∞tai (comets) and we long-haired stars ..., then the doubling of the sun ..., the fear of which has brought people to suspect that some celestial and divine power exists.

This cause together with the fourth – the orderly motion of the heavens and heavenly bodies (N.D. II 15) – is very similar to what Lucretius himself cites as the cause of mankind’s mistaken belief in intervening gods (V 1183-93). It is this misconception Lucretius sets out to dispel in book VI (as he promises in VI 48-90), pointing out that all these phenomena can be explained physically. Although diseases do not properly belong to meteorology, yet, just like meteorological phenomena, they frighten us, by our inability to understand their causes, into believing that they are brought about by the gods, and to eradicate this fear diseases need to be physically accounted for no less than meteorological phenomena and local marvels.298

2.3.3 Conclusion The inclusion of local marvels in works of meteorology, especially in sections concerned with terrestrial phenomena, is not exceptional. Except for the summer flooding of the Nile, which is treated on a par with other meteorological phenomena, local marvels are dealt with in meteorology in much the same way as in paradoxographical literature, to which they properly belong and from which they are probably drawn: in most meteorological works such local marvels as are dealt with are simply enumerated in long lists, with no, or hardly any, effort to explain them individually. Lucretius’ treatment of local marvels differs from that in other meteorologies in two important ways. In the first place, instead of the long lists found in some of the other meteorologies, Lucretius offers just a small selection. In the second place, each of the marvels he has selected is provided with an extensive explanation. In addition to the virtual absence of such explanations in the extant meteorologies, there are several considerations to suggest that explaining local 298

In his commentary on Cic. ND II 14 ‘pestilentia’ Pease (1958) seems to view the inclusion of diseases in Lucr. DRN VI in the same light.



marvels was a relatively new affair. Firstly, excepting the problem of the Nile flood, local marvels are not dealt with in Aëtius’ doxography, as they probably would have been if there had been a tradition of explaining them, resulting in divergent opinions. Secondly, in the case of local marvels Lucretius himself most often provides just one explanation, whereas problems of a more general nature, like thunderbolts and earthquakes, are accounted for with a number of alternative explanations,299 which are most often drawn from earlier accounts (see §1.4 above); this suggests that in the case of local marvels not many explanations had been devised before, upon which Lucretius could have drawn. It would be interesting to know whether this new approach to marvellous phenomena is due to Lucretius himself or perhaps to Epicurus or an intermediary writer. Again there are two considerations to suggest that the account of marvellous phenomena does not derive from Epicurus. In the first place, local marvels are not discussed in Epicurus’ Letter to Pythocles, nor is there any fragment or testimony to suggest that he discussed such phenomena elsewhere, except for magnets and diseases.300 In the second place, if – as I have suggested above – the passages on local marvels in Lucretius book VI are ultimately derived from paradoxographical works, and if – as is usually assumed – the paradoxographical genre was inaugurated by Callimachus (ca. 310-240),301 then it is chronologically improbable (though not impossible) for Epicurus (341-270) to have written on these subjects. If this is true, the passage must have been conceived either by Lucretius himself or by another Epicurean from whom Lucretius subsequently borrowed it.302 Lucretius or his source may have been inspired to do so by the fact that other meteorologies too incorporated them, yet without explanations. He may have felt himself entitled to do so by Epicurus’ own occasional exhortations to the reader to find out certain things for himself (Hdt. 45, 68, 83; cf. Lucr. DRN I 402-409; 1114-17 and VI 527-534): a good Epicurean is expected to apply the principles of Epicurean physics to other, as yet unsolved problems. In this respect Lucretius’ discussion of local marvels is the obverse of DRN VI 527534. There Lucretius chooses not to discuss a number of phenomena which Epicurus had discussed (snow, wind, hail, hoar-frost and ice), instead inviting 299

For an overview of the number of alternative explanations per problem in DRN V and VI see APPENDIX B on p.245 below. 300 See p.98 above. 301 On Callimachus as the founding father of paradoxography see Ziegler (1949) col. 1140; Schepens & Delcroix (1996) 383; Wenskus (2000) col. 311 with references. 302 Robert D. Brown (1982), 349, thinks that Lucretius may have been inspired by Callimachus.


the reader to find out by himself; here he includes a number of phenomena which Epicurus had not discussed, using for himself Epicurus’ explicit invitation to do so.

2.4 Order of subjects 2.4.1 Introduction Above I have compared DNR VI with a number of other meteorologies regarding the range and subdivision of their subject matter. It is now time to have a look at some other aspects as well. It has often been observed that some of the afore-mentioned works exhibit an especially close correspondence in the order of their subjects. This is the case with Lucretius VI, Aëtius III, and the Syriac meteorology, and to a lesser degree Epicurus’ Letter to Pythocles (from chapter 17 [99] onwards). Several scholars have produced useful synopses to bring out these similarities.303 Yet, most of these synopses suffer from lack of perspicacity and detail, and from a certain bias in their presentation, exaggerating the similarities by omitting some of the evidence, rather than letting the evidence speak for itself. Most of them also fail to include Epicurus’ Letter to Pythocles, which one would expect to be very close to Lucretius VI as well. I will therefore repeat the exercise in some more detail, paying due attention to all the resemblances as well as differences in the order of presentation of each of the four works. Incidentally, these same works also resemble each other in another respect. In all four of them, meteorological problems are generally accounted for by a number of different explanations. In Aëtius’ Placita every single explanation is attributed by name to one or several earlier thinkers, while Lucretius, Epicurus and the Syriac meteorology instead present them, without reference to their original authors, as equally possible alternative explanations. The subject of multiple explanations has been explored in Chapter One of the present work. Yet, this further similarity enforces the impression that these four texts are somehow more closely related to each other than to other writings of this genre. In this section, however, the focus will be on the similarities in the order of subjects.

2.4.2 The table In the table below the subjects of each of the four works are presented, in the order in which they occur, in four parallel columns. Subjects which are anomalously placed in comparison to the other works are printed in bold letters, while subjects that are merely appended to, or included in, another or a 303

Reitzenstein (1924) 34-5; Runia (1997a) 97; Sedley (1998a) 158. See also the lists in Kidd (1992) 305-6.



more general subject are bracketed and italicized. The subjects of the Letter to Pythocles are indicated by the chapter numbers of the edition of Bollack and Laks, which I have occasionally subdivided into an A and a B part; the traditional numbering is added between square brackets.

EPICURUS & LUCRETIUS AND THE SCOPE AND STRUCTURE OF ANCIENT METEOROLOGY 117 Table 2-4. The order of subjects in Aëtius, the Syriac meteorology, Lucretius and Epicurus Aëtius Placita III + IV 1

The Syriac meteorology

Lucretius DRN VI

1. milky way 2. comets and shooting stars

Epicurus Letter to Pythocles 17-31 + 35 17A [99]: clouds 17B [99-100]: rain

1. thunder

96-159: thunder

18 [100]: thunder

2. lightning

160-218: lightning

19 [101-2]: lightning

20 [102-3]: why lightning precedes thunder 21 [103-4]: thunderbolts

7. clouds

(164-172: why lightning precedes thunder) 219-422: thunderbolts (379-422: theological excursus) 423-450: whirlwinds (prēstēres) 451-494: clouds

8. rain

495-523: rain

3. thunder without lightning 3. thunder, lightning, thunderbolts and whirlwinds (typhōnes and prēstēres)

4. clouds, rain, snow and hail

4. lightning without thunder 5. why lightning precedes thunder 6. thunderbolts

9. snow

22 [104-5]: whirlwinds (prēstēres) 23 [105-6]: earthquakes 24 [106]: subterranean winds 25 [106-7]: hail

10. hail

26 [107-8]: snow

11. dew

27A [108]: dew

12. hoar-frost

27B [109]: hoar-frost 28 [109]: ice

5. rainbow

524-526: rainbow

29 [109-10]: rainbow 30 [110-1]: halo round the moon

6. rods and mock suns 7. winds

8. winter and summer 9-14. THE EARTH 15. earthquakes 16. origin and bitterness of the sea 17. ebb and flood 18. halo IV 1. the flooding of the Nile

Stob. 39 ‘On waters’ (*)

13. wind 13.43-54: whirlwinds (prēstēres) 14.1-13: halo round the moon 14.14-29: theological excursus 15. earthquakes

527-534: snow, winds, hail, hoar-frost and ice

31 [111]: comets 35 [114-5]: shooting stars

535-607: earthquakes 608-638: constant size of the sea 639-702: the Etna 712-737: the flooding of the Nile 738-839: poisonous exhalations 840-905: wells and springs 906-1089: the magnet 1090-1286: diseases

(* On the possibility that Stobaeus 39 goes back to a lost Aëtian chapter see the argument on p.80 above.)



2.4.3 Some observations The order of subjects of each of the four works (including Epicurus’ Letter to Pythocles) is so similar that one can hardly escape the impression that they must ultimately derive, as far as this order is concerned, from one and the same work. None of the four works, or so I believe, has entirely preserved this original order, but each one deviates from it as a result of conscious decisions made by their individual authors as well as unconsious mistakes made in the course of each text’s transmission. Below I shall try to indicate which (conscious and accidental) moves may have produced the order of subjects now present in each work, and what the original order may have been. Milky way, comets and shooting stars Working his way down through the phenomena of the Aristotelian sublunary world, Aëtius starts his account of tå metãrsia with the phenomena that Aristotle had assigned to the topmost part of the sublunary sphere just below the realm of the stars: the Milky Way and comets and shooting stars. All three subjects are absent from the Syriac meteorology and DRN VI. Two of them, shooting stars and comets, are discussed in Epicurus’ Letter to Pythocles, but only at the end, where they straddle a number of undisputably astronomical subjects. This suggests that Epicurus may have assigned comets and shooting stars to astronomy rather than meteorology (see p.87 above), and the same consideration may explain why Lucretius and the Syriac meteorology do not include them. Theological excursus The final part of Lucretius’ account of thunderbolts is an argument against the view that thunderbolts come from the gods. The argument consists of a series of rhetorical questions like: why do thunderbolts strike high mountains, why do they fall in uninhabited regions, why do they sometimes strike good, god-fearing people, and leave the evil-doers alone, etc.,304 all leading up to the inevitable conclusion that the falling of thunderbolts cannot be attributed to the gods. A very similar argument is also found in the Syriac meteorology, in a passage generally referred to as the ‘theological excursus’.305 Here too the core of the argument is a series of rhetorical questions all concerned with thunderbolts, and in most cases closely matching the rhetorical questions and 304

Mansfeld (1992a), 320, points out that the argument as such is an old one: see Aristoph. Nubes 398-402. See also n.282 on p.106 above, and text thereto. 305 See Daiber (1992) 280-1, Mansfeld (1992a), Van Raalte (2003) – who rejects the passage as a later interpolation.


their development in Lucretius. At the end of the excursus, however, the author rather oddly concludes that (14.25-6) “it is thus not right to say hurricanes306 [sic!] that they come from God”, even though the entire argument is about thunderbolts, not hurricanes. Most likely, therefore, ‘hurricanes’ is just a mistake for ‘thunderbolts’. Equally strange is the fact that the excursus does not, as in Lucretius, follow the exposition of thunderbolts (ch.6), but is appended to the chapter on haloes (ch.14), to which it does not apply at all. Daiber therefore suggests that the passage “actually [belongs] to the chapters on thunderbolts (ch.6) and on eÔrow and prēstēr (13.33-54)”.307 That it should belong to the chapter on thunderbolts seems obvious, but Daiber’s reference to eÔrow and prēstēr was probably elicited by the mention of hurricanes in the conclusion of the excursus, which – I argued – is just a mistake for ‘thunderbolts’. Mansfeld, ignoring Daiber’s mention of eÔrow and prēstēr, interprets his words as implying that the excursus “should probably be reallocated to [...] the chapter on thunderbolts”. Mansfeld himself opts for a different solution. In the introduction to the excursus its subject is stated as (14.14) “the thunderbolt” and “anything that has been mentioned”. This means that the excursus, given its present location after the discussion of the halo but before earthquakes, seems to apply to haloes (among other things), but not to earthquakes, which is strange because haloes are quite harmless and earthquakes are not. Mansfeld therefore conjectures that the excursus would originally have been the concluding chapter of the whole treatise, which would make the backward reference apply to earthquakes as well.308 Of these two proposals for reallocation of the excursus I prefer the first one. Mansfeld’s proposal assumes that the excursus was meant to apply to earthquakes as well. Yet, the excursus does not say anything to this effect. Instead its entire argument is about thunderbolts, and therefore the excursus would have been most naturally placed directly following the physical account of thunderbolts, just as the corresponding passage in Lucretius. Another parallel for this position is provided by Seneca, who incorporates a similar passage (NQ II 42ff) in his overall account of thunder and lightning.


Arabic ‫[ ﺍﻝﺯﻭﺍﺏﻉ‬az-zawābi‘], plural of ‫[ ﺍﻝﺯﻭﺏﻉﺓ‬az-zauba‘a], ‘storm’, ‘hurricane’, the same word used in 13.45, 47 and 51 as an alternate name for the prēstēr. 307 Daiber (1992) 280. 308 Mansfeld (1992a) 316 & 318.



Whirlwinds (prēstēres) The Syriac meteorology differs from the other three works in its placement of the discussion of the prēstēr.309 Whereas in the accounts of Aëtius, Lucretius and Epicurus the prēstēr is attached to thunder, lightning and thunderbolts, in the Syriac meteorology it is appended to the discussion of wind.310 In order to fully appreciate what is happening here it will be necessary to cast the net a bit wider, and include some other meteorological writings. In Aristotle’s Meteorology, and in most subsequent meteorologies, two types of whirlwind are distinguished: the typhōn and the prēstēr.311 Both types are closely linked to the thunderbolt, so much so that all three are considered manifestations of the same phenomenon, differing only by degree, the thunderbolt being wholly fiery, the prēstēr half, and the typhōn not at all. Aëtius conforms entirely to this tradition, discussing under one single heading not just thunder, lightning and thunderbolts, but also typhōnes and prēstēres.312 Theophrastus too can be placed in this tradition: in the De igne (1.8-9) prēstēres and thunderbolts are mentioned together, as examples of fire being produced through violent motion.313 In other words, for Theophrastus, as for Aristotle, prēstēres are fiery and closely related to thunderbolts.314 309

The relevant section is not preserved in the Syriac, but Bar Kepha’s Syriac paraphrase (194v a22, Daiber 188) has: SY+SYrP [prīsṭīs], an evident mistake, as Daiber notes, for rY+SYrP [prīsṭīr]. The Arabic translations have: ‫[ ﻑﺭﻱﺱﻁﻱﺭ‬frīsṭīr] (Bar Bahlūl 13.19, Daiber 208) and: ‫[ ﻑﺭﺱﻁﻱﺭ‬frĭsṭīr] (Ibn al-Khammār 13.43, Daiber 241). All these are obvious transliterations of the Greek prhstÆr [prēstēr], which in post-classical times was pronounced [prīstīr]. 310 As observed by Kidd (1992) 303, and Sedley (1998a) 159 &182. 311 Arist. Mete. III 1, 371a8-18; [Arist.] De mundo 4, 395a21-24; Arrian (fr.3. p.187.4-11 Roos-Wirth) apud Stob. Ecl. 29.2.4-7; Stoics apud Diog. Laërt. VII 154.4-6; Chrysippus SVF II 703 apud Aët. III 3.13; Seneca NQ V 13.1-3 (who calls the typhōn ‘turbo’, and defines the prēstēr as an ‘igneus turbo’, i.e. a fiery typhōn); Pliny NH II 133-134 (who mentions beside the turbo/typhōn and the prēstēr a number of other similar phenomena: vertex, procella, columna, aulon). 312 Cf. also Hesiodus Theog. 844-6: kaËma d' Íp' émfot°rvn kãtexen fioeid°a pÒnton / bront∞w te sterop∞w te purÒw t' épÚ to›o pel≈rou / prhstÆrvn én°mvn te keraunoË te fleg°yontow. – ‘And through the two of them {sc. Zeus and Typhoëus} heat took hold on the dark-blue sea, / through the thunder and lightning, and through the fire from the monster, / and the prēstēr-winds and blazing thunderbolt.’ (tr. Hugh G. EvelynWhite, slightly modified). 313 See however Theophrastus De ventis 53 where the prēstēr, without reference to its fiery nature, is said to be produced from the conflict of two opposed winds. 314 As observed by Kidd (1992) 303.


At first sight Epicurus and Lucretius may seem to belong to the same tradition, for they too discuss the prēstēr immediately following the thunderbolt. However, their account of the prēstēr differs from more traditional accounts in three important respects: (1) there is no mention of its fiery nature, (2) there is no reference to the typhōn as its less fiery counterpart,315 and (3) there is nothing in the text to suggest that the prēstēr and the thunderbolt (as well as the typhōn) are varieties of the same phenomenon. The same three observations also apply to the discussion in the Syriac meteorology, which closely matches the accounts of Epicurus and Lucretius. However, whereas Epicurus and Lucretius, while severing the traditional ties between the prēstēr and the thunderbolt, maintain the traditional order of subjects, the Syriac meteorology goes one step further and reassigns the prēstēr to the chapter on wind.316 It seems reasonable to assume that in this case the Syriac meteorology has departed from the original order of subjects, which has been preserved by Aëtius, Epicurus and Lucretius.317 Rainbow, halo, rods and mock suns As we observed above (see p.79), Aët. III 18 on the halo is out of place. In almost every other meteorology the halo is related to the rainbow and to rods and mock suns, wich in Aët. III make up chapters 5 and 6 respectively. It is likely, therefore, that Aëtius, too, discussed the halo contiguously to these two subjects. Jaap Mansfeld (2005) argues for the order rainbow – halo – rods & mock suns.318 This sequence is partly preserved in Epicurus’ Letter to Pythocles which discusses the rainbow and the halo in two subsequent chapters. In DRN VI and the Syriac meteorology only one of these phenomena is discussed: the rainbow (very briefly) in DRN VI, and the halo in the Syriac meteorology. Winds & earthquakes The subject of winds is differently placed in each of the four works. In Lucretius VI (527-534) wind is mentioned in an enumeration of subjects (snow, winds, hail, hoar-frost and ice) which the reader is invited to investigate for himself. This does not necessarily mean that Lucretius found all these subjects consecutively in whatever source he used; he has simply lumped them together at the end of his exposition of atmospherical 315

Lucr. VI 438 uses ‘turbo’ simply as a synonym for ‘prēstēr’. Sen. NQ V 13.1-3 and Plin. NH II 133-134, too, deal with whirlwinds in the context of winds, yet, like Aristotle and several others, they use the word ‘prēstēr’ in the limited sense of ‘fiery whirlwind’ (see n.311 above). 317 My conclusion is basically that of Sedley (1998a) 159 and 182. 318 See note 214 on p.79 above. 316



phenomena, because he has chosen not to discuss them. We should not, therefore, attach too much weight to the location of the subject of wind in Lucretius. Another problem presents itself in Epicurus’ Letter to Pythocles. Immediately following the chapter on earthquakes, Epicurus goes on to discuss winds (pneÊmata). There has been much debate about this chapter. Usener thinks it is a dislocated fragment of the account of prēstēres,319 and Bailey that it is a relic of a chapter on volcanoes (matching Lucr. VI 639-702), most of which would have been lost in a lacuna.320 Others prefer to see it as an explanation of the origin of atmospherical winds,321 thus providing the counterpart of Aëtius III 7 and the Syriac meteorology ch.13. However, there is no textual evidence to suggest a lacuna, and if the chapter had been about atmospherical winds, one would have expected Epicurus to use the normal Greek word for wind, ênemow, instead of pneËma, which may refer to any gust of air.322 The use of a definite article and a connecting particle at the beginning of the chapter (tå d¢ pneÊmata ...) in fact suggests that Epicurus, far from starting something new, is expanding on something he mentioned just before, viz. the subterranean wind (pneÊmatow §n tª gª) of the previous chapter, which he holds responsible for the production of earthquakes.323 Epicurus’ chapter 24, therefore, is probably not about atmospherical winds at all, and does not constitute a counterpart to Aët. III 7 and Syr. 13. The only drawback of this conclusion is that it robs the Letter to Pythocles of its chapter on atmospherical winds, which otherwise seems to have been a standard subject in ancient meteorology (see Table 2-1 on p.91 above). This leaves us with only two texts out of four, from which we might hope to learn something about the original position of the chapter on winds. Unfortunately the two texts diverge on this point: in the Syriac meteorology the halo is discussed after winds, but in Aëtius III the halo was probably discussed in connection with the rainbow and rods and mock suns before 319

Usener (1887) p.48 in apparatu critico ad loc.: “nam haec adduntur superiori de turbinibus loco” 320 Bailey (1926) 310; id. (1947) 1655. 321 Arrighetti (1973) 533. 322 In Aristotle’s Meteorology (I 13 & II 4-6) and the De mundo (IV 394b7-395a10) the word ênemow is used specifically to refer to atmospherical wind (cf. De mundo 4, 394b13: tå d¢ §n é°ri pn°onta pneÊmata kaloËmen én°mouw ...), while pneËma may refer to any (supposed) gust of air, including e.g. thunder, lightning, thunderbolts and whirlwinds (Mete. II 9 – III 1; De mundo 4, 395a11-24). Aëtius’ chapter on atmospherical winds (III 7) is called Per‹ én°mvn, as is Theophrastus’ treatise on winds. 323 Bollack & Laks (1978) 240-1.


winds (see previous heading). It is therefore impossible to say with certainty which order was original, but I will give it a try. Above (on p.94) we saw that Seneca and his fellow Stoics include earthquakes with atmospherical, rather than terrestrial, phenomena. The reason for this, Seneca states, is that earthquakes, being caused by air, should be dealt with in connection with other phenomena of the air. In this respect he closely follows Aristotle who attributes earthquakes to subterranean winds and discusses them immediately after atmospherical winds. It is probably because of this close connection that Diogenes Laërtius’ overview of Stoic meteorology, which is otherwise limited to atmospherical phenomena, also includes earthquakes. I also argued that the inclusion of earthquakes among ‘lofty’ phenomena in the Syriac meteorology and Epicurus’ Letter to Pythocles must be due to these works going back to this same tradition, even though they no longer endorse its underlying assumption that earthquakes are uniquely caused by winds. However, if both works include earthquakes with atmospherical phenomena because earthquakes were traditionally connected with winds, it seems likely that in the original order of subjects the section on earthquakes would also have immediately followed the chapter on winds (as in Aristotle’s Meteorology). If this is true the most likely original sequence of subjects would have been: rainbow – halo – rods & mock suns – winds – earthquakes, which is precisely the sequence we find in Aëtius III (accepting the reallocation of the chapter on the halo as proposed above). There is only one problem: in Aëtius III the chapters on winds (7) and on earthquakes (15) are assigned to different classes of phenomena, viz. metãrsia and prÒsgeia, and separated from each other by no less than seven intervening chapters, most of which concern the earth as a whole. By including these subjects in his meteorology Aëtius III deviates, not just from the three works mentioned above, but from almost every other work on meteorology (see p.93 above). Therefore the inclusion of these chapters by Aëtius most likely reflects a deliberate departure from his sources. Giving priority to the earthquakes’ location (earth) rather than their cause (wind/air), he chose to place the dividing line between atmospherical and terrestrial phenomena right between winds and earthquakes. Similar considerations may also have led him to move a number of subjects concerned with the earth as a whole from cosmology and astronomy – to which most cosmologists assign them – to the beginning of the terrestrial section. Another chapter, on the alternation of the seasons, he then placed at the end of the section on atmospherical phenomena, i.e. after the chapter on winds. If this reconstruction is correct, the rainbow will originally have been dealt with (as it still is in Aëtius III and Ep. Pyth.) right after the section on clouds, rain, snow, hail, etc., of which the rainbow is a possible



symptom. In the Syriac meteorology, on the other hand, the rainbow is omitted and the account of the halo moved to a position after the chapter on wind.324 Terrestrial subject other than earthquakes While in the Syriac meteorology and Epicurus’ Letter to Pythocles earthquakes are the only ‘terrestrial’ phenomena to be dealt with, Aëtius and Lucretius discuss several more: Table 2-5. Terrestrial subjects other than earthquakes in Aëtius and Lucretius Aëtius Placita III 16 – IV 1 16. origin and bitterness of the sea 17. ebb and flood IV 1. the flooding of the Nile (Stobaeus 39 ‘On waters’?) -

Lucretius DRN VI 608-1286 608-638: constant size of the sea 639-702: the Etna 712-737: the flooding of the Nile 738-839: poisonous exhalations 840-905: wells and springs 906-1089: the magnet 1090-1286: diseases

At first sight the correspondence between the two works, as far as the terrestrial phenomena are concerned, seems minimal: only one subject, the summer flooding of the Nile, is identical. If we stand back a bit, however, two further similarities become visible: in both works the account of the Nile flood is preceded by a section or sections dealing with the sea (even though the precise subjects are not the same), and – if we accept the ultimate Aëtian origin of the subject of Stobaeus Ecl. Phys. 39 ‘On waters’ (see p.80 above) – both works have a section or sections dealing with ‘waters’ (even though the precise subjects are not the same). If we take into account the large degree of correspondence in the order of subjects in the atmospherical sections of both works, the correspondence in the terrestrial section can hardly be a matter of coincidence.

2.4.4 The original order of subjects The table below presents a reconstruction of the original order of subjects that underlies both Aëtius III, the Syriac meteorology, Epicurus’ Letter to Pythocles 17ff and Lucretius DRN VI. Subjects whose inclusion in the original 324

Sedley (1998a), 159, also believes the Syriac meteorology’s chapter on the halo to be misplaced, on the grounds that in its present location it spoils the top-down sequence of phenomena.


work is uncertain have been bracketed; three dots indicate that other subjects may have been dealt with in between; subjects that are subordinate to others are indented, and subjects that seem to be misplaced in each of the four works are printed in bold. It must be born in mind that subjects which are represented by only one or two of the four extant texts may be later additions and not belong to the original sequence of subjects. Table 2-6. Proposed original order of subjects Aet. III ATMOSPHERICAL PHENOMENA Milky way Comets Shooting stars Thunder Lightning (Thunder without lightning) (Lightning without thunder) (Why lightning precedes thunder) Thunderbolts (Theological excursus) Whirlwinds (prēstēres) Clouds Rain Snow Hail Dew Hoar-frost (Ice) Rainbow Halo round the moon Rods and mock suns Winds Earthquakes TERRESTRIAL PHENOMENA … Subjects pertaining to the sea … The flooding of the Nile … (Other terrestrial waters) …

1 2 2 3 3

3 3 4 4 4 4

5 18 6 7 15

Syr. met. Ep. Pyth. Lucr. VI

1 2 3 4 5 6 14.14-29 13.43-54 7 8 9 10 11 12 14.1-13 13 15

31 35 18 19 20 21 22 17A 17B 26 25 27A 27B 28 29 30


96-159 160-218 164-172 219-422 379-422 423-450 451-494 495-523 527-534 527-534 527-534 527-534 524-526 527-534 535-607



IV 1


(Stob. 39)


2.4.5 Deviations from the original order Each of the four works deviates from the conjectured original order of subjects in its own way. Below I will briefly discuss the most important



omissions, additions and transpositions each of the four works may have undergone. Aëtius III: The original order of subjects seems best preserved by Aëtius. The only discrepancy is the chapter on haloes, which is illogically placed in the section on terrestrial phenomena, probably due to a scribal error. Aëtius’ account further differs from the other works by transferring the chapter on earthquakes from atmospherical to terrestrial phenomena and inserting a number of other chapters before, viz. one on the alternation of the seasons, and several dealing with the earth as a whole, subjects that were usually discussed in the context of cosmology and astronomy. Both additions should probably be ascribed to Aëtius’ wish to organize his subject matter strictly according to each phenomenon’s location. Syriac meteorology: The milky way, comets & shooting stars are not discussed in the Syriac meteorology. Their absence may indicate that the author did not consider these phenomena atmospherical, as Aristotle had done, but astronomical. Also absent is the rainbow, which is a standard ingredient in every other ancient meteorology. Its absence is even more conspicuous because the less well known halo, which is traditionally related to the rainbow, is dealt with. I cannot imagine any reason why the rainbow should be omitted, except for an accident in the text’s transmission. Three passages, moreover, seem to have been moved to a different position. The section on whirlwinds (prēstēres) has been transplanted from its traditional place next to thunderbolts to a new position at the end of the chapter on winds. This move may have been prompted by the author’s, the translator’s or a scribe’s realisation that whirlwinds are in fact winds and should be discussed in connection with these (see p.120 above). Two other moves are harder to account for. The halo, which, as we have seen, was originally dealt with in connection with the rainbow, has probably been moved from a position before, to a position after winds. Moreover, a theological excursus, in which the divine provenance of thunderbolts is refuted, has been separated from the physical account of thunderbolts, to which it obviously belongs, and been appended to the chapter on the halo, with which it has nothing to do. Perhaps these two apparently irrational transpositions and the omission of the rainbow can be accounted for on the assumption that a scribe first omitted to copy these passages, and later, when realising the omissions, inserted two of the missing passages (the halo and the theological excursus) at the point which he had reached by then, unfortunately still forgetting to include the rainbow.325 325

The dislocation of Aët. III 18 on the halo is explained similarly in Mansfeld (1995) 27 and Mansfeld & Runia (2009a) 44.


Lucretius DRN VI: Like the Syriac meteorology and possibly for the same reason (see above) Lucretius omits the Milky Way, comets and shooting stars. He further skips snow, hail, dew, hoar-frost and ice – perhaps so as not to bore his readers with accounts that would for the most part be only variations of what was already said about clouds and rain. He goes on to discuss the rainbow very briefly (524-526), but omits the halo and winds. Before moving on from lofty to earthly phenomena, he first sums up most of the phenomena he earlier omitted (527-534) – snow, winds, hail, hoar-frost and ice – inviting the reader to examine the possible explanations for himself. Epicurus Letter to Pythocles: The order of subjects in the Letter to Pythocles is so strange and irregular that the correspondence with the other three works is easily overlooked.326 The strangest feature is the return, after a long intermezzo on atmospherical phenomena, to astronomy. This unexpected return also makes it hard to know for certain whether Epicurus meant comets and shooting stars to go with astronomy or meteorology, although the evidence seems slightly in favour of the first option (see p.87 above). As for the truly meteorological part of the Letter, i.e. chs.17-30, although the order of its subject-matter may seem very different from the order found in the other three works, on closer inspection the differences amount to just two major transpositions. First, for some unknown reason clouds and rain have been detached from the other types of precipitation and placed before the account of thunder, lightning, thunderbolts and the prēstēr. Secondly, earthquakes and subterranean winds have been moved from their original place at the end of the passage, after the rainbow and the halo, to a position immediately following the account of the prēstēr, perhaps because of their equally destructive effects.327 Only after the chapters on earthquakes and subterranean winds does Epicurus return to the other kinds of precipitation. Atmospherical winds are omitted altogether. Although these transpositions have much disturbed the Letter’s order of subjects, two long sequences out of the original order remain virtually intact: thunder – lightning – thunderbolts – prēstēr; and (with the minor transposition of snow and hail): hail – snow – dew – hoarfrost – ice – rainbow – halo. Especially significant is the fact that in the first sequence the prēstēr has retained its traditional position after the thunderbolt, even though – as we have seen – in Epicurus’ view the two phenomena are not related (see p.120 above). 326

Runia (1997a) 97, comparing Lucr. VI and Aët. III, and Sedley (1998a) 157-8, comparing Lucr. VI, Aët. III and the Syriac meteorology, both ignore the order of subjects of the Letter to Pythocles. 327 Note that the overview of Stoic meteorology in Diog. Laërt. VII 151-154, whose order of subjects is even more garbled than Epicurus’, also jumps from prēstēres to earthquakes.



It appears to be possible to reconstruct an original order of subjects from which all four works derive, each of them deviating from it in its own special way. It would be interesting to know what work this order of subjects originally came from and in what way our four texts relate to this original and to each other. In order to come closer to answering these questions it will be useful to have a closer look at the structure of our four texts on the level of individual chapters and sections.

2.4.6 The internal structure of chapters and sections Although the order in which the different subjects are presented in each of the four works is very similar, the internal structure of chapters and sections is not always the same. A first point of difference can be gleaned from Table 2-4, on p.117 above. While Epicurus, Lucretius and the Syriac meteorology all tend to deal with each phenomenon in isolation, Aëtius sometimes collects a number of related subjects into one chapter. Chapter 3, for instance, deals with thunder, lightning, thunderbolts, prēstēres and typhōnes, which are accounted for by a number of integrated theories, and chapter 4 deals with clouds, rain, snow and hail in the same way. Of the three remaining works Epicurus’ exhibits the simplest structure. The account of almost every phenomenon is reduced to the question of its causation, which is accounted for with a list of alternative theories, sometimes followed by a brief methodological remark. In a few cases other aspects of the phenomenon under investigation are dealt with separately: in this way the account of lightning in general (ch.19) is followed by an account of why lightning precedes thunder (ch.20); the account of earthquakes (ch.23) by an account of the subterranean winds responsible for earthquakes; the account of hail (ch.25) by an account of the round shape of hailstones (ch.25a), the account of the rainbow (ch.29) by an account of its round shape (ch.29a); and the account of the halo (ch.30) by an account of the circumstances that may lead to a halo (ch.30b). A more complex structure is found in the Syriac meteorology (see APPENDIX C on p.246 below). Although some of its chapters are limited, as in Epicurus’ Letter to Pythocles, to listing a number of alternative explanations for a certain phenomenon’s occurrence, most of them deal with several related questions and aspects as well. The most outspoken examples are chapter 6 on thunderbolts, and chapter 13 on winds. A very similar structure, though less explicit (as there are no chapters), underlies Lucretius’ meteorology, which in its more complex sections deals with many of the same questions, and in a


somewhat similar order, as the Syriac meteorology. In the table below I have printed in parallel columns the contents of the Syriac’s and Lucretius’ sections on thunderbolts. Items that have no match in the other text are italicized. Table 2-7. Syriac meteorology chapter 6 and Lucretius VI 219-422 on thunderbolts Syriac meteorology chapter 6 2-9: The nature of thunderbolts 10-16: Their subtlety and penetration and speed 16-21: Causes of thunderbolts 21-28: Necessary conditions 28-36: Their escape from the cloud 36-67: Reasons for their downward movement 67-74: Why they are more frequent in spring 74-85: Why they are more frequent in high places 85-91: Their effects (discussed in 14.14-29)

Lucretius DRN VI 219-422 219-224: The nature of thunderbolts 225-238: Their subtlety and penetration 239-245: Introduction to the causes of thunderbolts 246-268: Necessary conditions 269-280: Causes of thunderbolts 281-294: Their escape from the cloud 295-322: More causes of thunderbolts 323-347: Reasons for their speed 348-356: Their effects 357-378: Why they are more frequent in autumn and spring 379-422: Thunderbolts not the work of the gods

Although the corresponding sections of the Syriac meteorology and DRN VI are structured very similarly, neither text can be reduced to the other: amidst matching items each text also includes items that have no counterpart in the other one. Similar observations can be made on the level of the individual items of each section. For instance, the last item of Lucretius’ section on thunderbolts, which deals with the question whether thunderbolts are instruments of the gods, offers virtually the same arguments as the corresponding section of the Syriac meteorology (14.14-29), but in a different order.328 Also the lists of alternative explanations correspond to a large degree – sometimes even including the analogies used to illustrate each explanation –,329 although the order in which they are presented often differs. As an example, I have printed below in two parallel columns the alternative explanations of thunder as offered in the Syriac meteorology and by Lucretius, with the corresponding illustrative analogies in square brackets:

328 329

Mansfeld (1992a) 326-7. The degree of corresponce in the illustrative analogies is variously assessed: Kidd (1992) 301 sees ‘close parallels including the illustrative analogies’ between Lucretius and the Syriac meteorolgy, while Garani (2007) 97 observes a ‘remarkable lack of correspondence between Theophrastean [i.e. in the Syriac meteorology] and Lucretian analogies.’



Table 2-8. The alternative explanations of thunder in the Syriac meteorology and Lucretius VI Syriac meteorology ch.1 2-5: (1) collision of concave clouds [clapping hands] 6-8: (2) wind whirling in hollow cloud [wind in caves and large jars] 9-11: (3) thunderbolt quenched in a moist cloud [white-hot iron quenched in cold water] 12-14: (4) wind hitting and breaking an icy cloud [flapping paper] 15-17: (5) wind blowing through crooked cloud [butchers blowing into guts] 18-20: (6) exploding cloud [explosion of inflated bladder] 21-23: (7) friction of clouds [millstones being rubbed together]

Lucretius DRN VI 96-159 96-115: (1) collision of clouds [flapping and tearing of a canvas awning] 116-120: (2) friction of clouds [-] 121-131: (3) exploding cloud [explosion of inflated bladder] 132-136: (4) wind blowing through ragged clouds [wind blowing through trees] 137-141: (5) wind rending cloud [wind uprooting trees] 142-144: (6) waves breaking in the clouds [waves breaking in rivers and the sea] 145-149: (7) thunderbolt quenched in a moist cloud [white-hot iron quenched in cold water] 150-155: (8) cloud burnt by thunderbolt [forest fire] 156-159: (9) frozen cloud breaking up [-]

Four explanations appear to be common to the Syriac meteorology and Lucretius, two of which are even illustrated by the same analogy (the exploding bladder and red-hot iron being quenched in water). Yet, the order in which these four explanations occur is entirely different for the two works. Comparison of other lists of alternative explanations will yield similar results, revealing a significant correspondence, which, however, rarely extends to the order of presentation. In sum: a comparison of the internal structure of the corresponding chapters and sections of our four meteorologies shows different degrees of similarity. While the Syriac and Lucretius’ meteorologies appear to be very similar, and hence probably closely related, Epicurus’ Letter to Pythocles is structured in a much simpler way, confining itself to the investigation of the causes of each phenomenon, which possibly testifies to the Letter’s summary nature. The chapters of Aëtius’ account, by contrast, are organized according to a different rationale, which may suggest that Aëtius’ book III is only distantly related to the other works.

2.5 Relations between the four texts Four meteorological texts, viz. Aëtius’ Placita III, Epicurus’ Letter to Pythocles (ch.17ff), the Syriac meteorology and Lucretius’ DRN VI, resemble each other closely in the order of their subjects, while the latter two show a large degree of correspondence in the structure of individual chapters and


sections as well. Although it is likely that these four texts are somehow related, it appears to be impossible to simply reduce them to each other. It is unlikely, therefore, that they are directly related, but the similarities in order and structure must have been transmitted by still other texts, which are no longer extant. In this section I will try to specify these missing links as best as the evidence allows. In order to do this it will also be necessary to deal with the identity of the Syriac meteorology, whose equation with (a part or a summary of) Theophrastus’ Metarsiology has hitherto been accepted too readily.

2.5.1 Epicurus’ Letter to Pythocles and his “other meteorology” In the Letter to Pythocles Epicurus claims to be merely summarizing what he wrote elsewhere (§n êlloiw; see also p.85ff above).330 As far as the cosmology and astronomy are concerned, these “other places” have been identified as (parts of) books XI and XII of the On nature. As for the meteorological portion of the Letter these “other places” have not yet been identified, but we can be certain that there existed some other, more elaborate, account of meteorology, whether this was part of book XII or XIII of the On nature, as Sedley suggests, or of some other of Epicurus’ numerous works. This “other meteorology”, as I shall call it, may well be the source of the reports on Epicurus’ seismology in Seneca’s NQ VI 20 and Aët. III 15.11, and on Epicurus’ views on rain and hail in Aët. III 4.5, which provide details not present in Epicurus’ Letter to Pythocles. In this “other meteorology” Epicurus will have discussed at least those subjects that are also dealt with in the meteorological section of the Letter (i.e. chs. 16-30, and perhaps also 31 and 35 on shooting stars and comets), and perhaps more (e.g. such subjects as are dealt with in the terrestrial parts of Aëtius III and Lucretius’ VI). Since the Letter to Pythocles is only a summary of Epicurus’ “other meteorology”, it is likely that the traces of the original order (see p.124ff. above) it still preserves were transmitted to it from this more extensive work.

2.5.2 Lucretius DRN VI and Epicurus’ “other meteorology” There has been much scholarly debate about Lucretius’ sources. Any investigation into these sources must begin with Lucretius’ own statement on the matter: in the whole body of the DRN Lucretius acknowledges only one source: the writings of Epicurus.331 I do not think this necessarily means that 330

Jaap Mansfeld points out to me, as he once did to David Sedley, that the plural §n êlloiw was sometimes used to refer to a single passage: see Sedley (1998a) 120 n.68. 331 DRN III 9-12: “... tuisque ex, inclute, chartis / floriferis ut apes in saltibus omnia libant / omnia nos itidem depascimur aurea dicta ...” – “… and from your pages, illustrious man,



Epicurus’ writings were Lucretius’ only source, but this is where any investigation should start.332 As for the sources of Lucretius’ meteorology in book VI the first work that comes to mind is, of course, Epicurus’ Letter to Pythocles, which deals with many of the same subjects. However, even a superficial comparison will show that the succint treatment in the Letter to Pythocles cannot have been the main source for Lucretius’ much richer account. Yet, as I have argued above, the Letter itself only summarizes what Epicurus had written on these matters elsewhere. Although this other, more detailed, account of meteorology is no longer extant, it is reasonable to assume that this was Lucretius’ main source, at least for the subjects up to and including earthquakes. Moreover, if both the meteorological part of Epicurus’ Letter to Pythocles and book VI of Lucretius’ DRN derive from Epicurus’ “other meteorology”, and if both works have each in their own way adapted the ‘original order of subjects’ established above, it seems likely that such traces of the original order as each of the two works preserves have come to them through Epicurus’ “other meteorology”. If so, Epicurus’ “other meteorology” must have been closer to the original order of subjects than either of the two works derived from it.

2.5.3 Authorship and identity of the Syriac meteorology It is now time to address a question I have been postponing for some time: is the Syriac meteorology, as it claims to be, a work by Theophrastus? The majority view nowadays seems to be that it is, that it is in fact a translation (of either the whole, or possibly a part) of Theophrastus’ Metarsiology,333 a work in two books of which otherwise only the title,334 a ‘table of contents’ (possibly incomplete),335 and a few paraphrasing fragments survive. 336

just as bees in the flowery woods sip everything, we likewise feed on all your golden words ...”. 332 So already Reitzenstein (1924) 37. 333 See n.166 on p.64 above. 334 Diog. Laërt. 5.44: Metarsiologik«n aÄ bÄ. 335 Proclus In Platonis Ti. 35A (II 121.3 Diehl = Theophr. fr.159 FHS&G): ... zhtoËntow, pÒyen m¢n afl bronta¤, pÒyen d¢ ênemoi, po›ai d¢ afit¤ai keraun«n, éstrap«n, prhstÆrvn, Íet«n, xiÒnow, xalãzhw, ì dØ kal«w poi«n §n tª t«n mete≈rvn afitiolog¤& (so Steinmetz (1964), 216-7, ms.: épolog¤&) t∞w prepoÊshw efikotolog¤aw ka‹ aÈtÚw ±j¤vsen ... – “... by investigating whence come thunders, whence winds, what are the causes of thunderbolts, lightnings, prēstēres, rains, snow, hail, which he too in his discussion of meteorological phenomena quite properly thought deserving of a fitting conjectural account ...”.


Another option, that the work goes back to Epicurus, was suggested and then swiftly rejected by Bergsträßer, the first editor of the first text found, and this procedure was repeated by several other early commentators.337 The most recent editor of all three versions, Hans Daiber, makes no reference to this option but confidently claims that the work is an unabridged translation of Theophrastus’ Metarsiology.338 I think there is still some reason for doubt. Above (§1.5.5 on p.64) I have already indicated that the systematic application of multiple explanations, such as we find in the Syriac meteorology, is very different from Theophrastus’ practice in other, undisputed, works. The best that recent commentators have been able to come up with is the observation that Theophrastus occasionally offers multiple explanations. Yet, as I have also shown, even in those cases Theophrastus usually offers far less explanations than the Syriac meteorology does, and generally does not support each alternative explanation with analogies from everyday experience, or derive his alternative explanations from the views of earlier thinkers. In all these respects the Syriac meteorology is closer to Epicurus’ Letter to Pythocles and the astronomical and meteorological passages of Lucretius DRN (V 509-770 and VI respectively), than to any undisputed work by Theophrastus. Significantly, Theophrastus was never in ancient literature cited as a champion of multiple explanations, as Epicurus was. On the contrary, one ancient witness, Seneca, explicitly claims that regarding the causes of earthquakes Theophrastus held the same, single, view as Aristotle,339 while citing Democritus and Epicurus for having offered multiple alternative explanations.340 The Syriac meteorology also differs from Theophrastus’ undisputed works in another respect. As we have seen above (p.120), the way prēstēres are viewed in the Syriac meteorology is very different from what we find in Theophrastus’ De igne, yet very similar to the corresponding passages in the works of Epicurus and Lucretius. In De igne 1.8-9 Theophrastus views the prēstēr, like Aristotle before him, as a fiery whirlwind, closely related to the thunderbolt. In the Syriac meteorology, on the other hand, prēstēr seems to be the generic word for whirlwind, which is nowhere said to be fiery or to be somehow related to the thunderbolt. The Syriac meteorology even goes one step further than Epicurus and Lucretius, by relocating the discussion of the 336

Theophrastus’ meteorological fragments have been collected in FHS&G (1992) 356-365 as frgs. 186A – 194. Of these only 186B and 192 are explicitly attributed to Theophrastus’ Metarsiology. 337 See n.172 and text thereto on p.65 above. 338 Daiber (1992) 285-6, 287. 339 Sen. NQ 6.13.1: “In hac sententia licet ponas Aristotelem et discipulum eius Theophrastum.” (= Theophrastus fr.195 FHS&G). 340 Sen. NQ 6.20 (= Democritus fr.A98 D-K / Epicurus fr.351 Us.)



prēstēr, from its traditional position after the thunderbolt, to the end of the section on winds. It might of course be argued that Theophrastus changed his view of prēstēres after writing his De igne and before composing his Metarsiology. That Theophrastus did in fact change his position might seem to be confirmed by a passage in his De ventis, which is supposed to have been written after the Metarsiology.341 Here, in ch.53, the origin of prēstēres is attributed, without any reference to their fiery nature or their connection with typhōnes and thunderbolts, to the conflict of contrary winds. However, the brief mention of prēstēres and thunderbolts in the De igne explicitly refers the reader to an earlier exposition of these phenomena, which is interpreted as a reference to the Metarsiology.342 Yet, if this is true, the Metarsiology must have presented the more traditional view of the prēstēr as a fiery whirlwind related to the thunderbolt, and not the innovative concept of a prēstēr as a special (and fireless) type of wind, as advanced by Epicurus, Lucretius and the Syriac meteorology, and perhaps also Theophrastus’ De ventis. These, to my view, are the most important objections against the identification of the Syriac meteorology with Theophrastus’ Metarsiology. None of them seems absolutely fatal: Theophrastus could have dealt with meteorological matters differently from other subjects (it will not do, however, to state that using multiple explanations was characteristic of him), Seneca could have been misinformed about Theophrastus’ account of earthquakes, and Theophrastus could have changed his view of prēstēres between writing his De igne and his Metarsiology (in which case the De igne must have been written earlier, and the supposed backward reference to the Metarsiology be explained in some other way). Yet, the effect of all these objections is cumulative, and shows that the alternative hypothesis of an Epicurean origin should at least have been taken more seriously. Yet, the assumption of an Epicurean origin is not without its difficulties either. Below I will discuss four passages that seem to testify to a Theophrastean (or at least Peripatetic) rather than an Epicurean origin. (1st) Chapter 6 of the Syriac meteorology, on thunderbolts, contains a section (36-67) concerning the downward motion of thunderbolts, which has no counterpart in Lucretius’ book VI. According to the Syriac meteorology (36-41), a thunderbolt reaches us (i.e. moves downward) either because winds beat the cloud on top, or because the cloud is split at the bottom. In both cases

341 342

Steinmetz (1964) 9 n1, 56; Daiber (1992) 286 Steinmetz (1964) 9 n2, 114; Daiber (1992) 273, 286.


the movement is the result of a force external to the thunderbolt.343 The fact that the Syriac meteorologist finds it necessary to explain the thunderbolt’s downward movement is significant. Although Lucretius does not deal with the subject in the corresponding portion of book VI, he refers to the question elsewhere. In DRN II 203-215 he warns his reader not to believe that things can move upward of their own accord. Everything material, including fire, has a natural tendency to move downward, a tendency which can only be checked or reversed by an external force. When, under normal circumstances, fire is seen to rise, this must be attributed to the surrounding air, which, by being heavier, squeezes the fire upwards. Therefore, when a thunderbolt (which is fiery) moves downward, this is due primarily to its own nature, and would seem to need no further explanation. However, it can be argued that even from an Epicurean perspective a further explanation may still be asked for. If, under normal circumstances, fire is seen to rise, even though this is not to be attributed to its own nature but to upward pressure from the surrounding air, one may reasonably ask why in the case of thunderbolts these normal circumstances do not apply. Therefore, an explanation of the downward movement of thunderbolts need not be un-Epicurean per se, even though it is absent from both Epicurus’ Letter to Pythocles and Lucretius’ book VI. However, the continuation of the explanation in the Syriac meteorology (41-48) is distinctly un-Epicurean: The reason that the cloud is split from the bottom and not from (42) the top is as follows: Those two vapors which ascend from the earth are joined, (43) namely the thick vapor and the fine vapor. If they ascend, (44) then the fine one of both (kinds) moves quickly upwards, because it approaches its natural place. (45) And that is because each one of the bodies, when it is distant (46) from its (natural) place, has a weak and slow movement; but if it is near to its (natural) place, (47) (its movement) is quick and strong. Therefore, whenever the fine vapor (48) ascends, it has a much quicker344 movement. (tr. Daiber (1992), modified)

Two elements in this passage bear witness to its Peripatetic origins: the theory of the two vapors and the notion of a natural place. The two vapors,


Cf. Arist. Mete. I 4, 342a13-27 on natural vs. forced motion of shooting stars (and thunderbolts), and II 9, 369a20-30 on the downward motion of thunderbolts, fallwinds (eknephiai) etc. 344 Daiber (1992), following the Arabic version of (presumably) Ibn al-Khammār, has ‘quick’, but Wagner, in Wagner & Steinmetz (1964), who follows the Syriac version, has ‘much quicker’ (‘viel schneller’). Unfortunately, the section in missing in the Arabic version of Bar Bahlūl.



which play a major part in Aristotle’s Meteorology,345 do not seem to be essentially incompatible with Epicurus’ atomism, but in fact both Epicurus, in the Letter to Pythocles, and Lucretius, in DRN VI, refrain from using them. The notion of a natural place is, however, typical of the Peripatos and entirely alien to Epicureanism. According to Aristotle, the sublunary sphere is subject to two opposing natural motions: heavy bodies move naturally downwards and light ones upwards; both natural motions have their end-point in a natural place: the downward motion ends in the absolute down which is the centre of the (finite) universe, and the upward motion ends in the absolute up which is the circumference of the sublunary sphere.346 Both motions accelerate when the natural place is approached.347 According to Epicurus, however, there is only one natural motion, viz. downward, to which there is no limit, as the universe itself is unlimited, and which is uniform (i.e. not accelerated).348 The notion of a natural place is therefore entirely unEpicurean. We may conclude therefore that the source of this passage, at least, cannot have been Epicurus. Still, there is something strange about the passage. While here the formation of clouds is attributed to the interaction between the two vapors, the next chapter (ch.7), which deals specifically with cloud formation, says nothing about the two vapors or exhalations. Could it be that the present passage and the chapter on clouds have different origins? (2nd) The two vapors, which we first encountered in ch.6 on thunderbolts, make a second appearence in ch.13 on winds. In fact the whole chapter, except for its final section on prēstēres (43-54), has a decidedly Peripatetic flavour. Unfortunately, we are not able to compare it with the Epicurean position, since both Epicurus and Lucretius fail to include a section on atmospherical winds in their meteorological overviews (see p.121 above). (3rd) Although, as we just observed, ch.7 on clouds does not mention the Peripatetic theory of the two vapors, in another respect it is quite Theophrastean. It is worth quoting the relevant lines (2-9 & 27-29): (2) The clouds come into existence for two causes: because of the accumulation (3) and thickness of air and its transformation into a watery 345

Arist. Mete. I 4, 341b6-13 et passim. Cf. also Phys. IV 1, 208b9-10 and 210a3-6; Cael. IV 3, 310a30-35. 346 Arist. Cael. IV (esp. chs.3-5 = 302a-304b). 347 Arist. Cael. I 8, 277a28-29 & b5-9; ibid. II 6, 288a17-22. 348 Epic. Hdt. 60-1.

EPICURUS & LUCRETIUS AND THE SCOPE AND STRUCTURE OF ANCIENT METEOROLOGY 137 substance349 or because of much vapor (4) which ascends and with which the ascending vapors of the seas as well as the remaining (5) fluids become mixed. Air comes together and becomes thick for two reasons: (6) because of coldness or because of contrary winds which squeeze it and bring it together. (7) We can observe something similar amongst us: When ascending vapor in (8) the bath encounters the roof and cannot penetrate this because of its thickness, (9) it accumulates and becomes water. {....} The clouds (28) turn into water, when they become very thick; their thickness is (29) caused by the pressure of hard winds or by coldness. (tr. Daiber (1992), modified)

In this case we are in a position to compare this with the views elsewhere attributed to Theophrastus. In his commentary on Aristotle’s Meteorology Olympiodorus writes 350: One should know, that while Aristotle says that cooling is the only cause of clouds turning into water, Theophrastus says that cooling is not the only cause of the production of water, but also compression. Note, for instance, that in Aethiopia, where there is no cooling, rain still pours down because of compression: for he says that there are very high mountains over there, agains which the clouds collide, and that subsequently rain pours down, because of the resulting compression. But in cauldrons too, says he, moisture runs down again, and in the vaults of baths, where there is no cooling; this clearly coming about through compression.

The text of the Syriac meteorology agrees with this report in two ways: firstly they both name the same two causes of water-formation in clouds, viz. cooling and compression, and secondly, they have one example in common, viz. the example of water-formation against the roof of a bath-house.351


The Arabic version of (presumably) Ibn Al-Khammār has ‘the nature of water’. According to Daiber (p.219) this is probably a misinterpretation of the Syriac (352b29), ‘a watery substance’. Unfortunately, the section is missing in the Arabic version of Bar Bahlūl. 350 Olympiodorus In Arist. Mete. I 9, 346b30 (p.80.30-81.1 Stüve) = fr.211B FHS&G (for the Greek text see p.63 above). See also Proclus In Plat. Tim. 22E (= Theophr. fr.211A FHS&G); Galen In Hippocr. Aer. 8.6 (= fr.211C ibid.), and Theophrastus De ventis 5.15. 351 Based on these correspondences Drossaart Lulofs (1955), 442, and Steinmetz (1964), 55, identify the two passages and then use Olympiodorus’ testimony as a proof of the incompleteness of the Syriac meteorology. Daiber (1992), 276 & 283-4, on the other hand, suggests that Olympiodorus may have added material from Theophrastus’ Per‹ Ídãtvn.



Again, neither explanation – cooling or compression – seems to be essentially incompatible with Epicurus’ atomism, but the fact is that for some reason both Epicurus and Lucretius, while including compression among a number of alternative causes,352 entirely ignore cooling.353 In this respect, then, the Syriac meteorology is closer to Theophrastus than to Epicurus and Lucretius. (4th) The second half of ch.14 is devoted to a refutation of divine interference in the case of thunderbolts. As was first observed by Jaap Mansfeld, the passage is very close in content and structure to Lucretius VI 379-422.354 Both passages offer a list of rhetorical questions (why do thunderbolts strike high mountains, why do they fall in uninhabited regions, why do they sometimes strike good, god-fearing people, and leave alone the evil-doers, etc.), all leading up to the conclusion that the falling of thunderbolts cannot be attributed to god or the gods. Lucretius leaves this conclusion pretty much unsaid, but in the Syriac meteorology the passage is introduced as follows: (14) Neither the thunderbolt (pl.) nor anything that has been mentioned has its origin in God. For it is (15) not correct (to say) that God should be the cause of disorder in the world; nay, (He is) the cause (16) of its arrangement and order. And that is why we ascribe its arrangement and order to God (17) [mighty and exalted is He!] and the disorder of the world to the nature of the world. (tr. Daiber (1992))

The first part of this remark is not unlike Epicurus. The sentence: “For it is not correct to say that God should be the cause of disorder in the world” is, as Mansfeld observes,355 very close to what Epicurus says in his Letter to Menoeceus, 134: “for nothing is done by a god in a disorderly way” (oÈy¢n går étãktvw ye“ prãttetai). However, what the Syriac meteorologist goes on to say is very un-Epicurean. Whereas the Syriac leaves god in charge of everything orderly in the world, Epicurus explicitly denies the gods’ 352

Epicurus on cloud formation (Pyth. 17A): N°fh dÊnatai g¤nesyai ka‹ sun¤stasyai (1) ka‹ parå pilÆseiw é°row pneumãtvn sun≈sei ...., and on the production of rain (Pyth. 17B): ÖHdh d' ép' aÈt«n (1) √ m¢n ylibom°nvn .... Lucretius on cloud formation (VI 463-64) ‘venti / portantes cogunt ad summa cacumina montis’, and on the production of rain (VI 510-512): ‘nam vis venti contrudit et ipsa / copia nimborum turba maiore coacta / urget et e supero premit ac facit effluere imbris’. 353 Also observed by Montserrat & Navarro (1991) 301 with n.78. 354 Mansfeld (1992a), 326-7. 355 Mansfeld (1992a), 325-6.


involvement in orderly and disorderly matters alike.356 The Syriac meteorology repeats its un-Epicurean position in the concluding part of the excursus (25-29). In sum, the following aspects of the Syriac meteorology seem to exclude an Epicurean origin of the text: (i) Section 6.36-67 accounts for the downward motion of thunderbolts, which in Epicurean physics does not need to be accounted for. (ii) Section 6.41-48 appeals to the Peripatetic theory of natural place which is incompatible with Epicurean physics. (iii) Section 6.41-48 and chapter 13 appeal to the Peripatetic theory of the two vapors, which is never found in Epicurus and Lucretius. In the first of these two passages the two vapors are invoked to account for the coming-intobeing of clouds. Surprisingly, in ch.7 which is entirely devoted to the subject of clouds the two vapors are not mentioned at all. (iv) In chapter 7 (lines 6 and 29) the Syriac meteorology ascribes cloud formation and rain production to cooling and compression, just as Theophrastus does in fr.211B FHS&G. Yet, cooling does not feature among the several explanations offered by Epicurus and Lucretius. (v) In the theological excursus (14.14-29) the Syriac meteorologist, like Epicurus and Lucretius, denies God’s responsibility for disorder in the world, but, unlike Epicurus and Lucretius, leaves God in charge of everything orderly (lines 15-17 and 27-29). Two of these objections (ii and v) are positively fatal to the assumption of an Epicurean origin of the Syriac meteorology (or at least the pertinent portions of it), a conclusion strengthened by the three remaining objections. An Epicurean origin of the entire document must therefore be rejected. It is still possible, however, that the Syriac meteorology is a compendium of some sort, derived for the most part from Epicurus’ meteorology, but supplemented and ‘corrected’ on the basis of other, possibly Peripatetic or even specifically Theophrastean theories. This hypothesis would account for the overal Epicurean, rather than Theophrastean, character of the work, and also for the curious fact that those features which object most strongly to an Epicurean origin are concentrated in just a few passages. It would seem, then, that we are left with two possible scenario’s: either the Syriac meteorology is, as the communis opinio would have it, a version of Theophrastus’ lost Metarsiology (allowance made for a certain amount of shortening, omissions, transpositions and perhaps also additions), or it is a 356

Epic. Hdt. 76-77; Lucr. DRN II 1090-1104; ibid. V 1183-93; 1204-10; Cic. ND I 52.



compendium largely, but not solely, dependent on Epicurean meteorology, into which certain peripatetic elements have been incorporated. How such a compendium came to be transmitted under the name of Theophrastus I could not say, but there is certainly nothing exceptional about such false ascriptions.

2.5.4 Lucretius, Epicurus and the Syriac meteorology As we have seen above, Lucretius DRN VI 96-607 and the Syriac meteorology resemble each other to such a degree, that it is hard to avoid the conclusion that they must be closely related. Yet, we have also seen that the relevant portion of DRN VI probably derives directly from a more extensive account of meteorology by Epicurus, which Epicurus himself later summarized in his Letter to Pythocles. If this is true, the relationship between Lucretius VI and the Syriac meteorology (whichever view we take of this work’s identity) must run via Epicurus’ more extensive meteorology. In the previous section two hypotheses about the status of the Syriac meteorology have been proposed. If the Syriac meteorology is, as is commonly believed, a (possibly shortened and garbled) version of Theophrastus’ Metarsiology, then this work of Theophrastus’ is likely to have been the immediate source for Epicurus’ more extensive meteorology,357 and thus an indirect source for Lucretius DRN VI 96-607 and Epicurus’ Letter to Pythocles 17-30. If, on the other hand, the Syriac meteorology is a compendium largely based on Epicurean meteorology, its most likely source is Epicurus’ more extensive account of meteorology, on which DRN VI 96-607 and Epicurus’ Letter to Pythocles 17-30 depend as well. In this scenario Theophrastus’ Metarsiology, which without the evidence of the Syriac meteorology we know nothing about, has no place. The two possible scenarios for the relations between Epicurus, Lucretius and the Syriac meteorology are presented below. Texts that are no longer extant are bracketed; arrows indicate the influence these works may have exercised on each other with respect to their structure and order of subjects (they are not meant to exclude the possibility of other, external, sources for particular problems and theories).


Sedley (1998a), 182, while acknowledging the close similarity between Lucretius VI and the Syriac meteorology, which he believes to be Theophrastus’ Metarsiology, strangely maintains that not this work but the corresponding section of Theophrastus’ Physical Opinions was Epicurus’ and therefore Lucretius’ ultimate principal source.

EPICURUS & LUCRETIUS AND THE SCOPE AND STRUCTURE OF ANCIENT METEOROLOGY 141 Figure 2-1: Possible relations between Epicurus, Lucretius and the Syriac meteorology Scenario 1:

Scenario 2:

2.5.5 Aëtius’ Placita and Theophrastus’ Physical Opinions This leaves us with only one more work whose relation to the rest needs to established. Above we have seen that the original sequence of subjects which seems to underly both Aëtius III and the three works mentioned above, is best preserved by Aëtius. He alone, like Aristotle before him, includes the Milky way, comets and shooting stars among atmospherical phenomena, and he alone preserves the original, Aristotelian, sequence wind - earthquakes (even though several chapters on the earth as a whole have been interposed). Yet, it is also clear, and not just for chronological reasons, that Aëtius III could not have been the original source from which the other three works derive their order and structure. It is more probable that it derives its order from the same original source as the other three works, but independently. An attractive candidate for this common origin would be Theophrastus’ Physical opinions,358 a doxographical work from which Aëtius’ Placita is commonly believed to be ultimately derived,359 and from which Epicurus is often assumed to have culled most of his alternative explanations in astronomy and meteorology (see §1.4 on p.53ff above).360


On Theophrastus’ work being called Fusika‹ dÒjai (Physical opinions) rather than Fusik«n dÒjai (Opinions of the physicists) see Mansfeld (1990) 3057-8 n.1and id. (1992b) 64-5. 359 Diels (1879) 102ff. For an overview of Diels’ views on the matter see e.g. Burnet (1892) 33-6; Regenbogen (1940) cols. 1535-9; Mansfeld & Runia (1997) 78-9. 360 For Epicurus’ dependence on Theophrastus’ Physical opinions see Usener (1887) xl-xli. Cf. also Sedley (1998a) 166-85.



It should be noted, however, that the evidence for Theophrastus’ Physical opinions is slight. Only a handful of fragments remain,361 which do not permit any conclusion about the work’s structure and order of subjects. If the Syriac meteorology really is, as is commonly believed, a version of Theophrastus’ Metarsiology, then we can safely assume that the corresponding portion of the Physical opinions, by the same author, had roughly the same order of subjects, which, therefore, it could have transmitted to Aëtius III. Yet, the attribution of the Syriac meteorology to Theophrastus is, as I have shown, by no means certain, and neither, therefore, is the structure and order of subjects of Theophrastus’ Physical opinions. Its relation to Aëtius’ Placita and Epicurus’ meteorological works is, therefore, at best conjectural,362 but it is the best hypothesis we have.

2.5.6 Summary Above we have explored the possible relationships between Epicurus’ Letter to Pythocles (17-30), Lucretius’ DRN VI (96-607) and the Syriac meteorology as regards their scope and order of subjects. We have also, tentatively, indicated how Aëtius III may relate to the other three works. It is now time to bring together the main threads of the argument and summarize our findings: 1. Both Lucretius’ DRN VI 96-607 and Epicurus’ Letter to Pythocles 17-30 derive their scope and order of subjects from a more extensive account of meteorological phenomena by Epicurus, now lost, that may have been part of book XII or XIII of his On nature. 2. The Syriac meteorology is either (scenario 1) a – possibly shortened and garbled –version of Theophrastus’ Metarsiology, which Epicurus all but reproduced as his own in his more extensive meteorological account, and from which he adopted the use of multiple explanations, which he subsequently extended to astronomical phenomena as well; or (scenario 2) it is a compendium largely based on Epicurus’ more extensive meteorological account, in which case the use of multiple explanations may well have been Epicurus’ own invention, and another source for the order of subjects and the individual explanations in his work must be assumed. 361

The collections of fragments made by Usener (1858); and Diels (1879) 473-95 are too inclusive: see e.g. Regenbogen (1940) 1536.68-1537.14; Steinmetz (1964) 334-351; Runia (1992) 116-7. 362 For a more cautious view concerning the relation between Aëtius’ Placita and Theophrastus’ Physical opinions see e.g. Mansfeld (1989) esp. 338-42; id. (1992b); id. (2005).


3. In scenario 1, Theophrastus’ Metarsiology is the most likely direct source of Epicurus’ more extensive meteorology, with Theophrastus’ Physical opinions as a possible secondary source.363 In scenario 2, there is no reason to assign any role to Theophrastus’ Metarsiology (which in this case we know nothing about) and Theophrastus’ Physical opinions may well have been Epicurus’ primary source. 4. In both scenarios, Theophrastus’ Physical opinions, from which Aëtius’ work is believed to be derived, is the most likely link between Aëtius on the one hand and Epicurus, Lucretius and the Syriac meteorology on the other. The two scenarios are illustrated below. Texts that are no longer extant are bracketed; arrows indicate the influence these works may have exercised on each other with respect to their structure and order of subjects, and dashed arrows indicate a possible alternative or additional influence. Figure 2-2: Possible relations between Aëtius, Epicurus, Lucretius and the Syriac meteorology Scenario 1:

Scenario 2:

With respect to the two scenarios presented above a number of remarks and reservations need to be made: 1. It must be borne in mind that the two schemas only indicate the major influences the works involved may have exercised on each other with respect to the scope and order of subjecs, and must not be read so as to exclude other, possibly external, sources for particular problems and theories. 2. There is one aspect which both of the above scenarios fail to explain. Whereas the Syriac meteorology and Epicurus’ Letter to Pythocles 17-30 restrict themselves to the explanation of atmospherical phenomena (and 363

See n.357 above.



earthquakes), Aëtius III and Lucretius’ DRN VI both continue their accounts with a number of terrestrial problems, one of which is common to both works, viz. the summer flooding of the Nile, while two other accounts, viz. those on the sea and on waters are at least thematically related (see Table 2-5 on p.124 above). This is sometimes taken as an indication that Lucretius drew directly upon the doxographical tradition as well (hence the dashed arrows in the illustrations above).364 This assumption, however, leaves the majority of the – mostly exceptional and local – phenomena in the latter part of Lucretius’ book VI unaccounted for. Their inclusion may have been, as I suggested above (§2.3 on p.99ff), a personal innovation by Lucretius in answer to the increasing popularity of such marvel stories in the paradoxographical as well as meteorological literature of his time. 3. It may be argued that the two scenarios are overly simplistic, and in a way they are: for as soon as we admit other sources for particular problems and theories beside the ones proposed above (see points 1 and 2), we must also admit the possibility that these other sources influenced the scope and order of subjects as well. On the other hand, although this possibility cannot be absolutely ruled out, it must also be observed that no other sources, beside the ones suggested above, are needed to account for the similarities in the scope and order of subjects of the four works involved.

2.6 Conclusions In this chapter I have compared Epicurus’ Letter to Pythocles and Lucretius’ DRN VI with a number of meteorological works both as to the range and subdivision of subjects included (§2.2) and to the order in which these subjects are dealt with (§2.4). The inclusion in the later part of DRN VI of a large number of local marvels also led me to make a comparison with the genre of paradoxography (§2.3), while the comparison of the order of subjects provided the occasion to examine the possible relations between the Letter to Pythocles and DRN VI and other meteorological texts, and to reexamine the claims about the authorship of the Syriac meteorology (§2.5). The most important findings of this chapter are the following. In §2.2: The range of subjects covered by the Syriac meteorology, which is the closest parallel to DRN VI, seems to be complete (except for the omission of the rainbow). Contrary to the majority of ancient meteorological texts, Lucr. DRN VI, Epic. Pyth. and the Syriac meteorology probably considered comets and shooting stars astronomical rather than meteorological 364

Cf. Runia (1997a), esp. pp.98-99.


phenomena. Along with the majority of ancient meteorological texts, Lucr. DRN VI, Epic. Pyth. and the Syriac meteorology assigned earthquakes to atmospherical rather than terrestrial phenomena. In §2.3: The later part of DRN VI differs from the earlier part as well as from corresponding sections of other meteorological works in predominantly discussing particular local phenomena (marvels). Although other meteorologies too refer to such phenomena, as a rule these are not themselves the objects of inquiry but only serve to convey some general point. The only notable exception is the summer flooding of the Nile which had since long been the object of physical speculation, and was sometimes included in a discussion of meteorological phenomena. However, particular local phenomena are more typically found in works of the literary genre known as paradoxography, which is not about physical inquiry at all, but simply contents itself with recounting marvellous stories. In DRN VI Lucretius deals with such phenomena in a way that is different from the approach in other meteorological as well as paradoxographical works, by making them the principal objects of inquiry rather than simply listing them. For chronological reasons it is unlikely that Lucretius’ treatment of these phenomena derives from Epicurus. It seems more likely that the passage on marvellous phenomena is an expansion by Lucretius himself. In §2.4: Four texts in particular, viz. Lucretius DRN VI, Epicurus’ Letter to Pythocles (17-30), the Syriac meteorology, and book III of Aëtius’ Placita, exhibit a remarkable similarity in the order of their subjects. It is possible to derive an original order, from which the order of each of these four texts is somehow derived. In §2.5: The many similarities in the range, subdivision and order of subjects in these four works suggest that they are somehow related. It is highly likely that the meteorological portions of Epicurus’ Letter to Pythocles and Lucretius’ DRN VI both go back to a more extensive account of meteorology by Epicurus, probably part of his magnum opus On nature. The relation this work bears to the Syriac meteorology depends on our views on its identity. If we accept its commonly accepted identification with Theophrastus’ Metarsiology, it is likely that this work was the main source for Epicurus’ meteorological writings. If, on the other hand, we take the – arguable – view that the Syriac meteorology is a compendium of mainly Epicurean meteorology, it will not tell us anything about Epicurus’ source, which instead we might identify with Theophrastus’ Physical opinions. The latter work, from which Aëtius doxographical work is supposed to be ultimately derived, may also provide the missing link between Aëtius and the three other works.

“Nous sommes en 50 avant Jésus-Christ. Toute la Gaule est occupée par les Romains... Toute? Non! Un village peuplé d’irréductibles Gaulois résiste encore et toujours à l’envahisseur.” - René Goscinny & Albert Uderzo, from the opening page of every consecutive album of Asterix (1961-present)

3 THE SHAPE OF THE EARTH 3.1 Introduction In Plato’s Phaedo (97d), which claims to report the conversations held by Socrates on the last day of his life in 399 BC, Socrates first tells us how, as a young man, about to start reading Anaxagoras, he expected to be told, among other things, whether the earth is flat or round. Later in the same dialogue (108e, 110b) Socrates states his present conviction that the earth is shaped like a sphere. While we cannot be certain that Socrates really thought any of these things, it is clear that Plato himself at least was convinced of the earth’s sphericity, as were most of his Greek contemporaries and those who came after; the last Greek philosopher reported to have advocated a flat earth is Democritus, who died around 370 BC. From this date onwards the earth’s sphericity, supported by an ever increasing amount of evidence, was accepted by all. Well, perhaps not all. It is often claimed that, in spite of all the evidence, Epicurus (341-270 BC) and his Roman follower Lucretius (99-55 BC) rejected the earth’s sphericity and stubbornly clung to the antiquated concept of a flat earth,365 a view that has elicited such qualifications as “singularly behind the 365

The Epicureans’ rejection of the earth’s sphericity or, conversely, their advocacy of a flat earth is stated for a fact by, for instance, Dreyer (1906) 171-2; Thomson (1948) 16768; Rist (1972) 47; Sedley (1976) 49; Schmidt (1990) 33, 215; Furley (1996) 119; Sedley (1998b) 346; Furley (1999) 420-1, 429; Milton (2002) 184; Chalmers (2009) 52; Sedley (2008) ‘Lucretius’ in The Stanford Encyclopedia of Philosophy (Fall 2008 Edition), [accessed 20 June 2010]; Konstan (2009) ‘Epicurus’ in The Stanford Encyclopedia of Philosophy (Spring 2009 Edition), [accessed 24 August 2010]; and the article ‘Flat Earth’ in Wikipedia, The Free Encyclopedia (12 June 2010, 15:38 UTC), [accessed 20 June 2010].



time”,366 “alarmingly retrograde”,367 “queerly indifferent to scientific truth”, 368 “scarcely [deserving] to be mentioned in the history of science”,366 “ludicrous” (a pun on the name ‘Lucretius’),369 and “sadly un-Greek”.368 By contrast, I know of only one serious attempt to defend the Epicureans.370 In a brief paper, published in 1996,371 David Furley sets out to answer three questions: 1. Why did the Epicureans hold on to the claim that the earth is flat? 2. Were they familiar with contemporary astronomy? 3. Did they know of its arguments and put up a reasoned defence of their own position? In answer to the first question Furley points out, firstly, that a flat earth follows inevitably from the Epicurean theory of atomic motion, and, secondly, that this shape also naturally commended itself to the Epicureans, who set such a high value on perception: the earth looks flat, and therefore we must start from the assumption that it is flat. With respect to the second question Furley refers to the fragments of Epicurus’ On nature, where Epicurus engages critically with the methods and pretentions of contemporary astronomy. In answer to the third question Furley argues that another controversional Epicurean theory, about the sun and the other heavenly bodies being as small as they appear to us, may have been devised in part in order to reconcile the Epicureans’ flat-earth cosmology with Aristotle’s observation that the aspect of the sky changes with latitude. Although this attempt to defend or at least to better understand Epicurus’ views must be applauded, Furley, like so many other scholars, omits to mention one important point: nowhere in Epicurus’ remaining works and fragments and nowhere in Lucretius’ De rerum natura do we find any explicit statement about the shape of the earth.372 What evidence there is, is at best circumstantial and not without some ambiguity. Even Epicurus’ ancient 366

Dreyer (1906), loc. cit. Thriceholy, ‘Flat Earth and the Ptolemaeic System’, [accessed 25 June 2010]. 368 Thomson (1948), loc. cit. 369 Ethical Atheist, The Flat Earth, ‘Chapter 5: Analysis of 7000 Years of Thinking Regarding Earth’s Shape’, [posted 26 June 2007, accessed 25 June 2010]. 370 Only recently I learned of the existence of an unpublished PhD thesis by Donald Paul Conroy (1976), which seems to cover much of the same ground as Furley’s article and the present chapter. Although I managed to get hold of a copy, it came to late for me to incorporate any of its findings. 371 Furley (1996). See also Furley (1999) 420-1, 429. 372 As is rightly observed by Munro (1864) vol. 2, p.341 ad DRN V 534; Woltjer (1877) p.123 ad DRN V 534ff; Ernout-Robin (1925-28) vol. 3, p.72 ad DRN V 534-536; Bailey (1947), vol. 3, p.1403 ad DRN V 534ff. Taub (2009) 114-5 (referring to Conroy (1976) 110: see n.370 above). 367



critics, who were always ready to make fun of him – e.g. for his claim that the sun is as big as it seems, or for seriously considering the possibility that the sun might be extinguished at night –, remain silent on this point. A more thorough investigation of Epicurus’ views concerning the shape of the earth should therefore start with a survey of the evidence for attributing a flat-earth cosmology to the Epicureans. In the commentaries this attribution is usually inferred from such passages as Epicurus Hdt. 60, Lucretius’ DRN I 1052ff and V 534-563, in which a parallel natural motion is argued for, or implied. Although it is hard to see how a parallel motion could be combined with anything other than a flat earth, it is methodologically wiser to distinguish such passages from those where the shape of the earth can be directly inferred from astronomical observations or theories. Another limitation of Furley’s paper is that it focuses exclusively on one piece of astronomical evidence, viz. the changing aspect of the sky when one moves to the north or to the south, while in fact several such proofs were known in antiquity. One wonders how Epicurus and Lucretius would or could have dealt with those. In this chapter I propose to conduct a thorough investigation into the views of the Epicureans concerning the shape of the earth, their motivation for these views and their attitude towards the relevant astronomical theories and proofs. Based on a large number of passages from Epicurus, Lucretius and other Epicureans, I will try to answer the following questions: 1. What natural motion did the Epicureans assign to the atoms and bodies in general? 2. Does this natural motion imply a flat earth? 3. Do the Epicurean views on astronomy presuppose a flat earth? 4. Why (if they did) did the Epicureans hold on to the claim that the earth is flat? 5. Were they familiar with contemporary astronomy? 6. Did they know of its arguments for the earth’s sphericity and put up a reasoned defence of their own position? I will go about this investigation as follows: first, in §3.2, I will deal with some preliminary issues that may serve as a background to the investigation. Then, in §3.3, I will discuss a number of passages from Epicurus, Lucretius and other Epicureans that are (or may seem to be) relevant to one or more of our questions. I will take my lead from Lucretius, for two reasons. In the first place Lucretius provides most of the evidence, while such evidence as is furnished by other sources is mostly parallel to what we find in Lucretius (and will be so presented). Secondly, following Lucretius allows us to read most of the relevant passages as part of one continuous account, which may provide additional clues as to the underlying argument. Finally, in section 3.4, I will



summarize and combine my findings and try to present a balanced answer to the questions formulated above.

3.2 Some preliminaries Before I go on to deal with the views of Epicurus and Lucretius themselves some preliminary information may be useful or enlightening. I will start, in §3.2.1, by giving a brief historical overview of ancient views concerning the shape of the earth. Then, in §3.2.2, I will briefly discuss the most important ancient proofs that were brought forward in favour of a spherical earth, and indicate whether they were, or might have been, known to Epicurus and Lucretius. In the following subsection, §3.2.3, I will briefly deal with Epicurus’ ancient critics and what they might tell us about Epicurus’ views concerning the shape of the earth. Finally, in §3.2.4, I will discuss how the shape of the earth relates to the direction of natural motion or ‘gravity’.

3.2.1 The shape of the earth in antiquity: a historical overview The world picture that arises from the earliest works of Greek literature, the epics of Homer and Hesiod, is fairly simple.373 The earth is a flat disk, encircled by the waters of Ocean, with the heavens as an inverted bowl above, and the Tartarus in a corresponding position below.374 The views of Thales of Miletus (624-547), the first philosopher, do not seem to have been all that different. According to Thales, the earth is a disk floating like a log or a vessel on a primordial sea.375 If our reports are true, the first important step towards a radically different world view was made by Thales’ younger compatriot Anaximander (610-546). While sticking to the image of a flat earth (a column drum or a cylinder),376 he dispensed with the need for an underprop, such as Thales’ water. Instead he claimed that the earth, being equably related to every portion of the surrounding heavenly sphere, has no reason to move in any direction, and therefore does not fall down.377 If this account of 373

On the history of ancient astronomy in general, see e.g. Heath (1913) 7-129; id. (1932) xi-lv; Dicks (1970), Evans (1998). On the shape of the earth in particular, see e.g. Thomson (1948) 94-122; Evans (1998) 47-53. 374 Dicks (1970) 29-30; Heath (1913) 7; id. (1932) xi; Furley (1989a) 14. Cf. also Geminus Isagoge 16, 28. 375 Arist. Cael. II 13, 294a28-32 [fr.A14 D-K] and Sen. NQ III 14 [fr.A15 D-K]. The ascription to Thales of a spherical earth in Aët. III 10, 1 cannot be correct, because (1) it is incompatible with its floating, (2) it is anachronistic with respect to his successors’ views, and (3) it is contradicted by the explicit claims that either Pythagoras or Parmenides was the first to make the earth spherical (see note 380 below). 376 Hippol. Ref. I 6.3 [fr.A11 D-K]; Aët. III 10, 2 [fr.B5 D-K]; Ps.-Plut. Strom. 2 [fr.A10 DK]. 377 Arist. Cael. II 13, 295b 10-16 [fr.A26 D-K]; Hippol. Ref. I 6, 3 [fr.A11 D-K].



Anaximander’s views is correct,378 he certainly did not convince his successor Anaximenes (584-526), who returned to the old view of an earth in need of support: according to Anaximenes, the earth is flat like a table, and rides upon the air on account of its flatness.379 It was at the other end of the Greek world, in southern Italy, that the idea first arose that the earth might be a sphere, an idea proposed by either Pythagoras (ca. 580-500) or Parmenides (ca. 500) – on this the sources are divided.380 We are not told what brought them to this idea, but they may have been inspired by Anaximander’s equilibrium theory, which, according to our reports, was also accepted by Parmenides.381 Perhaps they thought that if the earth was equably related to every portion of the surrounding heavenly sphere, the earth’s shape too should be so related, i.e. spherical. The idea of a spherical earth remained current in subsequent centuries among Pythagoras’ followers,382 and it is quite likely that the sphericity of the earth was also accepted by Empedocles (ca. 490-430), who is said to have been a follower of both Pythagoras and Parmenides,383 and who in one of his fragments testifies to a centrifocal conception of up and down (see §3.2.4 below).384 378

Furley (1989a), 17-22, is inclined to think that the ascription of this theory to Anaximander is anachronistic. 379 Arist. Cael. II 13, 294b 13-30 [fr.A20 D-K]; Hippol. Ref. I 7, 4 [fr.A7 D-K]; Ps.-Plut. Strom. 3 [fr.A6 D-K]; Aët. III 10.3 [fr.A20 D-K]; Aët. III 15.8 [fr.A20 D-K]. 380 Diog. Laërt. VIII 48 [Parmenides A44 D-K]. Cf. Diog. Laërt. VIII 25 [Pythagoristae A1a D-K], and IX 21 [Parmenides A1 D-K]. 381 Aët. III 15.7 [fr.A44 D-K]. 382 Arist. Cael. II 2, 285b22-27 ascribes to the Pythagoreans the view that the north pole and the inhabitants of the northern hemisphere are above, whereas the south pole and the inhabitants of the southern hemisphere are below, and Arist. Cael. II 13, 293b25-30 describes how the Pythagoreans defend their claim that the earth is not at the centre by pointing out that even on the supposition of a centrally placed earth, its inhabitants are not at the centre but half the diameter away from it, which nevertheless doesn’t seem to affect our observations of the heavens. Both these theories presuppose a spherical earth. See also Dicks (1970) 72-73. 383 Diog. Laërt. VIII 54-56. 384 Fragment B35 D-K, lines 3-4: [...] §pe‹ Ne›kow m¢n §n°rtaton ·keto b°nyow / d¤nhw, §n d¢ m°s˙ FilÒthw strofãliggi g°nhtai, [...] = ‘when Strife reached the lowest depth / of the vortex, and Love came to be in the middle of the whirl ...’ (‘the lowest depth’ and ‘the middle’ seem to indicate the same place), and lines 9-10: ˜ssÉ ¶ti Ne›kow ¶ruke metãrsion: oÈ går émemf°vw / t«n pçn §j°sthken §pÉ ¶sxata t°rmata kÊklou = ‘all the things which Strife retained up high: ‘for it had not (yet) altogether retreated perfectly from them to the outermost boundaries of the circle.’ (‘up high’ and ‘the outermost boundaries of the circle’ seem to indicate the same thing). On the other hand, in Aëtius II 8 concerning the inclination of the cosmic axis with respect to the flat earth,



In the meantime the eastern Greeks still adhered to the traditional view. The historian Herodotus (484-425) probably believed the earth to be flat,385 and Anaxagoras of Clazomenae (500-428) and Diogenes of Apollonia (ca. 460) both adopted Anaximenes’ view of a flat earth floating on air.386 They were followed in this respect by the atomists from Abdera, Leucippus (early 5th century) and Democritus (ca. 460-370).387 By this time, however, word of the new theory had reached Athens, where for a time both theories existed side by side. In Plato’s Phaedo (97d) we hear that young Socrates (470-399), upon taking in hand a volume of Anaxagoras’, expected to be told, among other things, whether the earth is flat or round. In the end the second view won the day. In the same dialogue (108e - 109a) Plato (427-347) has Socrates, much older now, explain that he has been convinced by someone that the earth is a sphere and remains where it is, because, being placed equably in the centre of heaven, it has no reason to move in this direction rather than that – the same combination of theories that we saw attributed to Parmenides. From then on the idea of a spherical earth spread rapidly. In fact, there is no explicit information of anyone later than Democritus propounding a flat-earth cosmology.388 Even within the circle of his followers, there seems to have been some dissent. If we may believe the scanty information provided by Diogenes Laërtius,389 a certain Bion (early fourth century BC?), who ‘was a follower of Democritus and a mathematician, from Abdera, {...} was the first to say that there are regions where the night lasts six months and the day six months.’ The only regions for which this statement applies are the north and south poles, and Bion could only have Empedocles is mentioned alongside the flat-earthers Anaxagoras and Diogenes of Apollonia (see p.225 with n.565 below). 385 Furley (1989a) points to Herodotus’ report in III 104 that in India the hottest time of the day is in the morning, and that from then on the temperature steadily drops until at sunset it is extremely cold. This, according to Furley, indicates a flat earth, where in the east the sun arrives vertically overhead very soon after sunrise. Another proof may be found in IV 42 where Herodotus, reporting on the circumnavigation of Africa, expresses disbelief at the sailors’ claim that during their westward passage around the southern portion of the continent they had the sun on their right, i.e. to the north. With the image of a spherical earth before him, Herodotus would have had no reason for doubting the sailors’ statement. 386 On Anaxagoras see Arist. Cael. II 13, 294b 13-30; Hippol. Ref. I 8, 3 [fr.A42 D-K]; Diog. Laërt. II 8 [fr.A1 D-K]; Exc. astron. cod. Vatic. 381 (ed. Maass Aratea p. 143) [fr.A87 D-K]. On Diogenes see Schol. in Basil. Marc. 58 [fr.A16c D-K] and Aët. II 8.1. 387 On Leucippus see Aët. III 10.4 [fr.A26 D-K]; Diog. Laërt. IX 30 [fr.A1 D-K]. On Democritus see Arist. Cael. II 13, 294b 13-30; Aët. III 10.5 [fr.A94 D-K]. 388 That is, until around 300 AD, when the Christian writer Lactantius (Div. Inst. III 24) rejected the spherical earth on account of its incompatibility with the Holy Scripture. 389 Diog. Laërt. IV 58.4-6. See Hultsch (1897) 485–487; Abel (1974) 1014.



arrived at such a claim on the basis of a firm understanding of all the implications of a spherical earth.390 Also Eudoxus (408-355), the leading mathematical astronomer of his day, probably accepted the earth’s sphericity. This much at least can be inferred from his theory about the Nile flood (which he generously attributed to the priests of Heliopolis), viz. that the Nile had its sources in the southern hemisphere where it is winter when we have summer,391 a theory that presupposes a spherical earth. Moreover, his famous theory of concentric spheres on which the fixed stars and planets move also seems to require a spherical earth at the centre of the cosmos. To Aristotle (384-322) we owe the first thorough argument concering the earth’s shape.392 Aristotle accounts for the earth’s sphericity and its stable position in the centre of the cosmos on the assumption of a general centripetal tendency of all heavy matter. He also offers two astronomical proofs for the earth’s sphericity – the circular shape of the earth’s shadow during a lunar eclipse, and the changes in the sky’s aspect when one moves to the north or the south –, and he informs us that mathematicians had calculated the earth’s circumference, arriving at a number of 400.000 stades (≈ 72.000 km). Not much later the voyages of Pytheas of Marseille (ca. 325 BC) in the northern Atlantic, and Alexander’s conquests of Egypt, Persia and parts of India (334323) provided the geographers and astronomers with a large body of new observations on which to base their theories and calculations. Especially Pytheas is important in this respect: he is the first person on record to have used the length of the shadow of the gnomon to determine latitude,393 and also the first to record the maximum daylength for a number of different latitudes (up to 24 hours in Thule on the arctic circle).394 This, then, was the state of 390

For the theoretical background of Bion’s claim see e.g. Achilles 35.23-38; Cleom. I 4, 219-231; Ptol. Alm. II 6, 116.21-117.9; Plin. NH II, 186-187. 391 Aët. IV 1.7 (Eudoxus fr.288 Lasserre). 392 Arist. Cael. II 14, 297a8 – 298b20; cf. also Mete. I 3, 340b35-36. 393 Strabo I 4, 4: ‘... for as to the ratio of the gnomon to its shadow, which Pytheas has given for Massilia, this same ratio Hipparchus says he observed at Byzantium, at the same time of the year as that mentioned by Pytheas.’ [transl. Horace Leonard Jones (1917), modified] & Strabo II 5, 8: ‘But if the parallel through Byzantium passes approximately through Massilia, as Hipparchus says on the testimony of Pytheas (Hipparchus says, namely, that in Byzantium the ratio of the gnomon to its shadow is the same as that which Pytheas gave for Massilia), ...’ [transl. Horace Leonard Jones, modified]. 394 Gem. Is. VI 9 states that Pytheas reported on places where the night was extremely short, only 2 or 3 hours; Plin. NH II 186-7 ascribes to Pytheas the incorrect view that in Thule (six days sayling north of Britain) the day and the night each last 6 months (this is in fact only true of the geographical north and south poles), but at NH IV 104,



affairs when, around 300 BC, Epicurus (341-270) devised his (presumably) flat-earth philosophy. But the story does not end here. At around the same time Zeno of Citium (333-264), whose followers – the Stoics – were to become the Epicureans’ most fervent adversaries, decided for a spherical earth.395 Not much later Eratosthenes of Cyrene (276-194) made his famous and remarkably accurate calculation of the earth’s circumference.396 Then, around 150 BC the Stoic scholar Crates of Mallos constructed the first terrestrial globe,397 while Hipparchus (ca. 190-129) prescribed using simultaneously observed eclipses to establish differences in longitude.398 Finally, during Lucretius’ own lifetime (ca. 95-55), Posidonius (135-51), another Stoic scholar, made a new, very influential, estimate of the earth’s circumference.399 Nor was familiarity with the sphericity of the earth restricted to the Greeks: among the Romans too the theory had found currency. Among Lucretius’ contemporaries both Varro (116-27) and Cicero (106-43) accepted the earth’s sphericity.400

apparently referring back to the previous passage but not mentioning Pytheas, he corrects his report, saying that in Thule at the summer solstice there is no night at all. Strabo II 5, 8 may be referring to the same observation when he states that according to Pytheas in Thule the arctic circle (i.e. the circle that comprises the ever visible stars) coincides with the summer tropic, a view that is astronomically equivalent to what Plin. NH IV 104 says. 395 Aët. III 10.1 [SVF II 648]; Diog. Laërt. VII 145.6-7 [SVF II 650]; Achilles Isagoge 4 [SVF II 555]; Cic. N.D. II 98.5-6 [not in SVF]. Cf. Diog. Laërt. VIII 48.10-12 [SVF I, 276] where Zeno claims Hesiod’s authority for his own theory. 396 Cleomedes I 7, 49-110. For other ancient reports as well as modern literature on this measurement see Bowen-Todd (2004) ad loc. 397 Strabo II 5, 10.3-5. 398 Strabo I 1, 12: ‘In like manner, we cannot accurately fix points that lie at varying distances from us, whether to the east or the west, except by a comparison of the eclipses of the sun and the moon. That, then, is what Hipparchus says on the subject.’ [transl. Horace Leonard Jones] 399 Cleomedes I 7, 7-48. For other ancient reports as well as modern literature on this measurement see Bowen-Todd (2004) ad loc. 400 Varro Men. fr.516 Bücheler, apud Non. 333, 25 ‘in terrae pila’; Cic. Rep. VI 15 = Somn. Scip. 3, 7. Cf. Cic. Tusc. I 68.12 and ND II 98.5-6.


CHAPTER THREE Figure 3-1: Time-line of ancient theories on the shape of the earth Flat earth

Spherical earth 800 BC

Homer Hesiod 700 BC

Thales Anaximander Anaximenes

600 BC

Anaxagoras Herodotus Leucippus Democritus

500 BC

400 BC

300 BC Epicurus ?

Pythagoras ? Parmenides Empedocles ? Pythagoreans Bion of Abdera ? Plato Eudoxus Aristotle Pytheas Zeno of Citium Eratosthenes

200 BC Crates of Mallos Hipparchus 100 BC Lucretius ?

Posidonius Cicero Varro

3.2.2 Ancient proofs of the earth’s sphericity In his De caelo II 14, 297b24-298a10, Aristotle provides two astronomical proofs of the earth’s sphericity. In subsequent centuries many more proofs were devised. Other lists of proofs are provided by the Roman poet Manilius (before 14 AD), the Roman encyclopedist Pliny the elder (23-79 AD), the Greek philosopher Theon of Smyrna (ca. 70-135 AD), the Greek astronomer Ptolemy (ca. 85-165 AD) and the Greek philosopher Cleomedes (ca. 200 AD). One proof is also reported by the Greek geographer Strabo (64 BC - 24



AD).401 The longest lists are those of Cleomedes (five proofs) and Pliny (six proofs). Below I will discuss each of the ancient proofs, briefly explaining how they work, who reported them, when they may have been devised, and how – if at all – they might have been refuted by someone wishing to uphold the assumption of a flat earth: 1. According to Aristotle, the convex shape of the earth’s shadow as it passes over the moon during a lunar eclipse proves that the earth must be spherical. In fact, however, such a shadow could be produced by many different forms, even by a flat, disk-shaped earth, as Furley observes.402 For this reason, perhaps, this ‘proof’ was not repeated by subsequent authors. The argument depends, of course, on the assumption that the moon receives its light from the sun, and is sometimes robbed of this light by the interposition of the earth. Another possible way to escape the consequences of this argument would therefore be to deny that lunar eclipses are produced in this way. 2. Another proof, mentioned by Strabo, Pliny, Theon, Ptolemy and Cleomedes, is based on observations of, and from, approaching and departing ships. If a ship approaches land, then from the ship the mountain-peaks are seen first, and then gradually the lower-lying portions of the land seem to rise from the water. In the same way, from the land the top of the mast is seen first, and then gradually the rest of the ship appears to rise from the water. And when the ship departs then conversely the land and the ship are seen to sink under water. These effects can only be explained on the assumption that the surface of the water is curved. The oldest known report of this proof is provided by Strabo, who postdates both Epicurus and Lucretius. It is possible, therefore, that Epicurus and Lucretius were not familiar with this proof. If they had been familiar with it, and had wished to refute it, they might have dismissed such observations as mirages of the kind that is often observed above water. 3. As Aristotle observed,403 and many others repeated afterwards, the position of the stars changes according to the observer’s latitude. Some stars that are seen in the south are invisible in more northerly countries, and some stars that are continuously visible in the north, are seen to set and rise in the south. Such observations are commonly explained on the assumption that the 401

The relevant texts and passages are given in Table 3-1 below. Furley (1999) 421. 403 Furley (1996), 121 (referring to Dicks (1970) pp.87-8, who does not say so, and Vlastos (1975) pp.38-40), claims that there is some evidence that this proof may have been known to Euctemon in the last half of the fifth century. Dicks (1970) n.380 (referring to Eudoxus frs.75a+b Lasserre) suggests that Eudoxus may have been Aristotle’s source for this proof. For our purposes it is enough to know that it would have been available to Epicurus. 402



earth’s surface is curved from north to south. David Furley (1996) suggests that observations of this kind can be reconciled with a flat earth, if the heavenly bodies are assumed to be relatively close by. He compares this to the effect of someone walking in a big room under a painted ceiling: as he walks he will observe the same paintings from different angles. In fact, however, this model only explains part of Aristotle’s observation: while it may account for the fact that observers at different latitudes see the same stars at different heights, it fails to explain how certain stars can pass out of sight altogether. 4. A special case of the previous proof, briefly mentioned by Pliny and discussed more fully by Cleomedes, concerns the height of the celestial north pole. Cleomedes observes that in northern countries, such as Britain, the celestial north pole is seen high above the northern horizon, but in southern regions, such as southern Egypt and Aethiopia,404 appears low in the sky. These observations, he claims, can only be accounted for on the assumption that the earth’s surface is curved from north to south. Although the use of such observations as proof of the earth’s sphericity cannot be dated to anyone before the time of Pliny the elder, the essential ingredients were known to Hipparchus (ca. 190-120 BC),405 and may well go back even further. Chronologically, therefore, there is no reason why Lucretius could not have known of the observations relating to this proof. The consequences of this proof are, however, easily dissipated on the assumption that the heavenly bodies, including the celestial north pole, are relatively close by, as Furley suggests. 5. Not just the position of the celestial north pole, but also the position of the sun at noon on a given day of the year varies with latitude. In our part of the world, the midday sun stands generally higher in southern countries, and lower in northern countries, and accordingly shadows are longer in the north and shorter in the south. This observation, which is another special case of proof 3, is presented as a further proof of the earth’s sphericity by Pliny the elder. The relevant observations are much older, however. Pytheas of Marseille (ca. 325 BC) already used the length of the shadow of the gnomon to determine latitude.406 Consequently, Epicurus and Lucretius could have known of these observations. However, just as with the previous proof, these 404

Text and translation are quoted on p.226 and p.229 below. In his In Arati et Eudoxi Phaen. I 3, 6-7, Hipparchus provides the average values for Greece of: (a) the ratio of the gnomon to its shadow at noon during the equinox (cf. proof 5), (b) the maximum day-length (cf. proof 6), and (c) the polar height (tÚ ¶jarma toË pÒlou) (cf. proof 4), indicating that these three values are different at other latitudes. 406 See note 393 above. 405



observations can be reconciled with a flat earth on the assumption that the heavenly bodies are relatively close by. 6. Another proof of the earth’s sphericity is based on the observation that the maximum daylength increases with increasing latitude. Although Pliny is the only author to present this observation as a formal proof of the earth’s sphericity, the facts were widely known.407 The first person on record to measure latitude by the maximum daylength was, again, Pytheas of Marseille.408 Chronologically, therefore, Epicurus and Lucretius could have known the facts which underly this proof. In this case, however, there is no easy way to escape its consequences. As Cleomedes already observed, on a flat earth sunrise and sunset would be the same for everyone, and accordingly there would be no latitudinal variation of daylength.409 7. The last proof I wish to address concerns the time difference between places lying at different longitudes. Pliny reports that on several occasions when warning-fires were alighted successively from west to east, the last one was observed to be alighted at a much later (local) hour than the first. (Pliny’s argument would have been more convincing if the fires had been alighted from east to west, and the last one been observed to be alighted at an earlier local time than the first.) However, a much more secure way to establish timedifferences between places consisted in simultaneous observations of lunar eclipses. Hipparchus had already prescribed this procedure as a means to accurately measure differences in longitude,410 and the same procedure was cited as a proof of the earth’s sphericity by Manilius, Pliny, Theon, Ptolemy and Cleomedes. Concrete examples are offered by Pliny and Ptolemy, who both refer to the lunar eclipse of September 20, 331 BC, which was observed at a certain hour near Arbela in Mesopotamia, several days before the famous battle, and at a much earlier hour in Carthage (so Ptolemy) and Sicily (so Pliny).411 Although the argument cannot be dated with certainty till before the time of Hipparchus, Pliny’s and Ptolemy’s reference to the eclipse of 331 BC suggests that the argument may have been known much earlier. Lucretius could have known it, and it could have been known to Epicurus himself. As with the previous proof it is hard if not impossible to avoid its consequences. The only way out, it would seem, is to contest the observations themselves. One might point to the fact that lunar eclipses are extremely rare, that


See e.g. Strabo II 5.38-42; Geminus Isag. VI 7-8, Cleom. II 1, 438-451; Ptol. Alm. II 6, Mart. Cap. VIII 877. 408 See note 394 above. 409 Concerning this proof see also subsection 3.2.3 below. 410 See note 398 above. 411 Pliny, loc. cit.; Ptol., Geogr. I 4, 2.



simultaneous observations from different locations are even rarer, and that the historical records (such as Pliny’s and Ptolemy’s) are often conflicting. The following table provides a summary of the ancient proofs of the earth’s sphericity, with reference to the authors, texts and passages reporting them, and their earliest datable occurrence: Table 3-1: Ancient proofs of the earth’s sphericity Aristotle Manilius Strabo Pliny Theon Smyrn. Ptolemy Cleomedes 384-322 BC before 14 AD 64 BC - 24 AD 23-79 AD ca. 70-135 ca. 85-165 ca. 200 ? Cael. II, 14 Astron. I Geogr. I, 1 N.H. II Expos. III Alm. I, 4 Cael. I, 5 1. Convexity of the earth’s shadow during a lunar eclipse 2. Observations of, and from, departing and approaching ships 3. Aspect of the sky varying with latitude

Aristotle 384-322 BC

297b24-31 20.18-27 297b31 298a10






Strabo 64 BC 24 AD

177.8 179.4


15.23 16.13


Aristotle 384-322 BC


Hipparchus ca. 190-120 BC

4. Elevation of the pole varying with latitude


5. Length of shadows varying with latitude


6. Max. day-length varying with latitude


7. Longitudinal time difference established by eclipses

Earliest datable occurrence



Pytheas ca. 325 BC


14.19 15.23


Pytheas ca. 325 BC


Hipparchus ca. 190-120 BC

Of the proofs mentioned above, the 1st, 3rd, 5th and 6th could chronologically have been known to Epicurus, while Lucretius could have known the 4th and 7th as well. If, as Furley suggests, the Epicureans did try to put up a reasoned defense of their own flat-earth cosmology against contemporary astronomy, one would expect them to betray a knowledge of, and engage with, the proofs the astronomers offered for the sphericity of the earth. In our survey of Epicurean passages in section 3.3 below, we will therefore also look for clues that might tell us whether Lucretius, Epicurus or other Epicureans were aware of these observations, and whether or how they managed to reconcile them with their own views concerning the shape of the earth.

3.2.3 Epicurus’ ancient critics In antiquity, Epicurus was criticized and even ridiculed for many of his doctrines. He was attacked, for instance, for assigning upward and downward



directions to the infinite universe,412 for assigning a random and uncaused motion, the par°gklisiw or ‘swerve’, to the atom,413 for holding that the sun is the size it appears to be,414 and for contemplating the possibility that the sun might be extinguished at sunset and rekindled at dawn.415 If Epicurus had also claimed, in contrast to everybody else, that the earth is flat, would his critics not have seized the opportunity to attack him for yet another ‘stupidity’? Yet, no such criticism has come down to us. One of Epicurus’ most fervent critics was Cleomedes, who in the second book of his astronomical treatise attacks Epicurus by name for two of his theories.416 Yet, in the first book, in a section where the theory that the earth might be flat is explicitly refuted, Cleomedes fails to name Epicurus. If he had known Epicurus to be a flat-earther, would he not have named him? And yet Cleomedes comes very close to actually identifying Epicurus here: in I 5.11-13 those who believe the earth to be flat are said to be ‘following only the sense presentation based on sight’, which is exactly how the Epicureans are characterized in II 1.2-5.417 What is more, one of Cleomedes’ actual arguments against the earth being flat might as well be aimed directly at the Epicureans. In I 5.30-37 Cleomedes argues that those who believe the earth to be flat must also believe that the sun rises and sets for everyone at the same time, which is observably wrong. However, in II 1.426-451 he ridicules Epicurus for believing that the sun may be extinguished at sunset and rekindled at sunrise, pointing out that on a spherical earth sunset and sunrise occur at different times in different regions. Cleomedes might at this point have concluded that Epicurus must be thinking of a flat earth, but he does not. Instead he opts for a reductio ad absurdum, suggesting that according to Epicurus’ theory the sun must be at the same time extinguished for some observers and rekindled for others – and this incalculably many times. Again one wonders why Cleomedes does not simply accuse Epicurus of claiming the earth to be flat. The answer must be that Cleomedes (or his source) could not find any explicit statement to this effect in Epicurus’ works. All the other criticisms, 412

Chrysippus (SVF II 539) apud Plutarch De Stoic. repugn. 44, 1054b. See also on p.195 below. 413 Cicero De fin. I 18-20 (Epic. 281a Us.); id. De fato 46 (Epic. 281c Us.). See also on p.194 below. 414 Cleomedes II 1.1-413; Cicero, De fin. I 20; Acad. II 82. See also on p.213 below. 415 Cleomedes II 1.426-466; Servius in Verg. Georg. I 247 (Epic. 346a Us.), id. in Verg. Aen. IV 584 (Epic. 346b Us.). See also on p.218 below. 416 Cleomedes II 1.1-413 on the size of the sun; II 1.426-466 on the sun’s daily extinction and rekindling. 417 Cleom. I 5.11: ... aÈtª tª katå tØn ˆcin fantas¤& katakolouyÆsantew ... and Cleom. II 1.2-5: ... aÈtª tª diå tØw ˆcevw fantas¤& katakolouyÆsantew ...



those of Cleomedes as well as other critics, were aimed at theories that could be explicitly ascribed to Epicurus. The view that the earth is flat, on the other hand, could not. Thus the critics’ failure to adress this subject is another indication, in addition to Lucretius’ silence on the matter, that Epicurus never explicitly opted for a flat earth. What we will be looking for, then, in our survey of relevant Epicurean passages, is circumstantial evidence, consisting in passages which either entail or presuppose a flat earth.

3.2.4 The direction of natural motion and the shape of the earth In his article ‘The Dynamics of the Earth’, David Furley distinguishes two fundamentally different ancient theories of natural motion. In one of these, which he calls linear or parallel, natural motion takes place along parallel lines. Up and down are defined in accordance with this parallel motion of fall, and the cosmos can be divided into an upper and a lower hemisphere. The tendency of heavy objects to fall down also applies to the earth as a whole, which, in order not to fall, must be supported by something underneath, be it water (Thales), or air (Anaximenes, Anaxagoras and Democritus), or the earth itself, extending downwards to infinity (Xenophanes).418 In the other theory, which Furley calls centrifocal, all natural motion is defined by the centre: down is motion towards the centre, up motion away from the centre, and the circular motions of the heavenly bodies are motions around the centre. The earth itself, being heavy, is also prone to move towards the centre, but being already agglomerated around the centre, it will not move; therefore there is no need for external support. This downward tendency applies to animals and human beings as well, allowing them to stand on every side of the earth, with their feet pointing towards the centre. People living diametrically opposite us are called antipodes, i.e. ‘having their feet pressed against ours’.419 Furley’s prototype of a centrifocal cosmology is the one propounded by Aristotle, who assigned three different natural motions to the elements: earth and water moving towards the centre, air and fire moving away from the centre, and aether or the first element (or the fifth as it is 418

Arist. Cael. II 13, 294b 13-30. See also Arist. Mete. II 7, 365a20-37, where the parallel cosmology of Anaxagoras is explicitly opposed to Aristotle’s own centrifocal system. 419 Cf. Cic. Luc. 123, 7: “Vos etiam dicitis esse e regione nobis, e contraria parte terrae qui adversis vestigiis stent contra nostra vestigia, quos antipodas vocatis.” See also Kaufmann (1894).



commonly called) moving in circles around the centre. Other centrifocal cosmologies are those of Plato, Strato and the Stoics. Strato, for instance, although he seems to have dispensed with Aristotle’s first element and assigned a natural downward motion to all four sublunary elements (attributing the apparent upward tendency of air and fire to extrusion by the heavier elements), still defined this natural downward motion as motion towards the centre.420 Plato’s cosmology differs from Aristotle’s in one important respect. In the Timaeus, 62c-63a, Plato observes that everyone speaks of ‘up’ and ‘down’ with respect to themselves, and that accordingly for someone standing on the other side of the earth ‘up’ and ‘down’ are exactly reversed. For this reason he rejects the use of these terms in a cosmological context altogether. In the De caelo, IV 1, 308a14-24, Aristotle critizises this view and attempts to save the terms ‘up’ and ‘down’ in an absolute sense by redefining them as ‘away from the centre’ and ‘towards the centre’ respectively. After Aristotle these definitions of ‘up’ and ‘down’ were accepted by most centrifocalists.421 This does not mean, however, that all parallel-linear terminology was banned from these centrifocal systems. When speaking of a star’s setting (dÊsiw / dusma¤) and rising (énatolÆ), or its height (Ïcow) or elevation (¶jarma), they were in fact, as we still are, using a parallel-linear spatial reference system, where ‘height’ is used to denote the (angular) distance from the plane of the horizon, and not the distance from the centre of the earth. Each of the two cosmological systems is also associated with one particular shape of the earth: in parallel cosmologies the earth is usually considered flat, while centrifocal theories are thought to imply a spherical earth. In the De caelo, Aristotle offers two arguments to demonstrate the truth of this implication. In the first place the all-sided centripetal pressure exerted by many individual chunks of earth, all seeking the centre of the universe, would naturally result in a spherical shape. Aristotle illustrates this by demonstrating that this shape would also have resulted if the earth had been generated at some time in the past (which he does not believe) (Cael. II 14, 297a12-b18). A similar argument was later put forward by the Stoics (Cic. ND II 116). In the second place a spherical shape logically follows from the observation that all heavy bodies fall at right angles to the surface of the earth, which in a centrifocal universe is only possible if the earth is spherical (Cael. II 14, 420 421

For the evidence on Strato see n.494 on p.188 below. For the Stoics, see e.g. Ar. Did. fr.31, apud Stob. Ecl. I, p. 184, 8 W. (= SVF II 527): tÚ m°son shme›on toË kÒsmou {...}, ˘ dØ toË pantÒw §sti kãtv, ênv d¢ tÚ épÉ aÈtoË efiw tÚ kÊklƒ pãnt˙, Cic. ND II 84: ‘in medium locum mundi, qui est infimus’, ibid. 116: ‘id autem medium infimum in sphaera est’, and esp. Cleomedes, I 1,158-192 (partly quoted as SVF II 557).



297b18-24). For the link between parallel natural motion and a flat earth no such arguments have come down to us, unless we count the view of Anaximenes, Anaxagoras and Democritus, who claimed that in order to be able to float on air the earth had to be flat and broad. However, a very plausible argument is provided by Furley, who points out that, as long as we accept the observed fact that heavy objects everywhere fall at right angles to the earth’s surface (cf. Aristotle’s second argument above), a parallel motion of fall requires a flat earth.422 The two cosmological systems with their implied shapes of the earth are illustrated below: Figure 3-2: Linear or parallel cosmology

Figure 3-3: Centrifocal cosmology

I have dealt with these two cosmological systems at some length because, as we shall see, the case for imputing a flat earth on the Epicureans largely depends on the assumed logical connection between the direction of natural motion and the shape of the earth. A clear understanding of the the two theories of natural motion and their assumed implications for the shape of the earth will be a useful tool in our survey of Epicurean passages.

3.3 Discussion of relevant passages After these preliminaries I will now turn to the promissed discussion of relevant Epicurean passages. In my search for such passages I have scanned through the whole of Lucretius’ DRN, Epicurus’ Letters to Herodotus and Pythocles, the fragments of book XI of Epicurus’ On nature, and the cosmo422

Furley (1976) 90; (1981) 10, 12; (1983) 91-2, 99; (1999) 42.



logical and astronomical fragments of Diogenes of Oenoanda’s inscription, as well as the commentaries thereto, looking for signs that might be, or have been, interpreted as somehow implying, or otherwise relating to, a specific shape of the earth. Lucretius appears to be the richest source of such passages, and for this reason I have chosen to take my lead from him. Such passages as can be found in the works and fragments of Epicurus and Diogenes are mostly parallel or supplementary to these, and will therefore be discussed in connection with them. In my discussion of these passages I will basically follow Lucretius’ order, which may help us to understand how the different passages hang together. I have also stumbled upon an interesting fragment of Philodemus which seems relevant to my purpose and has no exact parallel with Lucretius or other Epicureans. A discussion of this fragment will be appended to the list. The following passages will be dealt with: Table 3-2: Passages to be discussed Lucretius I 1052-93 II 62-250 IV 404-413 V 204-205 V 449-508 V 534-563 V 564-591 V 621-636 V 650-704 V 762-770 VI 1107 -

Parallel passages Epic. Hdt. 60 + fr.281 Us. Cic. ND I, 24; Philod. P. ye«n III 70 (Aëtius I 4) Epic. Nat. XI fr.42 Arr. Epic. Pyth. 6 [91] Diog. Oen. fr.13 I.11-13 Epic. Pyth. 7 [92] + 15 [98] Epic. Pyth. 13 [96] + scholion Philod. Sign. xxx 20-27

Subject Rejection of centrifocal natural motion Parallel downward motion Apparent proximity of the sun Climatic zones? Centrifocal cosmogony Stability of the earth Size of the sun Centrifocal terminology Sunrise and sunset The conical shadow of the earth The limp of the cosmic axis The varying length of shadows

Discussed in: 3.3.1 (p.163) 3.3.2 (p.190) 3.3.3 (p.198) 3.3.4 (p.199) 3.3.5 (p.202) 3.3.6 (p.213) 3.3.7 (p.213) 3.3.8 (p.217) 3.3.9 (p.218) 3.3.10 (p.219) 3.3.11 (p.222) 3.3.12 (p.230)

3.3.1 The rejection of centrifocal natural motion (DRN I 1052ff) Introduction The first passage that seems to be relevant to our subject, and one that is often cited in support of the claim that the Epicureans held the earth to be flat, is found at the end of the first book of DRN (1052ff). Here Lucretius rejects a theory of centrifocal natural motion, which, as we have seen, is generally assumed to imply a spherical earth (see p.160ff above). On the following pages I will analyse this passage, paying due attention to its context, to the intended target, to Lucretius’ refutation of the theory, and to the positive conclusions regarding Lucretius’ own views that may be drawn from the passage.


CHAPTER THREE Context of the passage The passage itself begins in line 1052, but in order to fully understand it we must first deal with the preceding passage. In line 951 Lucretius had broached a new subject: the infinity of the universe and of its two constituent parts: matter and empty space. As such Lucretius’ argument is an extended version of the more condensed argument in Epicurus’ Letter to Herodotus, 41-2: (i) ÉAllå mØn ka‹ tÚ pçn êpeirÒn §sti: tÚ går peperasm°non êkron ¶xei, tÚ d¢ êkron par' ßterÒn ti yevre›tai:
View more...


Copyright © 2017 DOCIT Inc.